2021
Calvo-Zaragoza, J.; Rizo, D.; Iñesta, J. M.
Reconocimiento Óptico de Partituras (OMR) aplicado al Fonde de Música Tradicional IMF-CSIC Book Chapter
In: Gambero-Ustárroz, M.; Ros-Fábregas, E. (Ed.): Musicología en Web. Patrimonio musical y Humanidades Digitales, Chapter 4, pp. 87-109, Edition Reichenberger, 2021, ISBN: 978-3-967280-14-2.
@inbook{k470,
title = {Reconocimiento Óptico de Partituras (OMR) aplicado al Fonde de Música Tradicional IMF-CSIC},
author = {J. Calvo-Zaragoza and D. Rizo and J. M. Iñesta},
editor = {M. Gambero-Ustárroz and E. Ros-Fábregas},
isbn = {978-3-967280-14-2},
year = {2021},
date = {2021-01-01},
booktitle = {Musicología en Web. Patrimonio musical y Humanidades Digitales},
pages = {87-109},
publisher = {Edition Reichenberger},
chapter = {4},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inbook}
}
Ortega-Bastida, J.; Gallego, A. J.; Rico-Juan, J. R.; Albarrán, P.
A multimodal approach for regional GDP prediction using social media activity and historical information Journal Article
In: Applied Soft Computing, pp. 107693, 2021, ISSN: 1568-4946.
@article{k471,
title = {A multimodal approach for regional GDP prediction using social media activity and historical information},
author = {J. Ortega-Bastida and A. J. Gallego and J. R. Rico-Juan and P. Albarrán},
issn = {1568-4946},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
journal = {Applied Soft Computing},
pages = {107693},
abstract = {This work proposes a multimodal approach with which to predict the regional Gross Domestic Product (GDP) by combining historical GDP values with the embodied information in Twitter messages concerning the current economic condition. This proposal is of great interest, since it delivers forecasts at higher frequencies than both the official statistics (published only annually at the regional level in Spain) and the existing unofficial quarterly predictions (which rely on economic indicators that are available only after months of delay). The proposed method is based on a two-stage architecture. In the first stage, a multi-task autoencoder is initially used to obtain a GDP-related representation of tweets, which are then filtered to remove outliers and to obtain the GDP prediction from the consensus of opinions. In a second stage, this result is combined with the historical GDP values of the region using a multimodal network. The method is evaluated in four different regions of Spain using the tweets written by the most relevant economists, politicians, newspapers and institutions in each one. The results show that our approach successfully learns the evolution of the GDP using only historical information and tweets, thus making it possible to provide earlier forecasts about the regional GDP. This method also makes it possible to establish which the most or least influential opinions regarding this prediction are. As an additional exercise, we have assessed how well our method predicted the effect of the COVID-19 pandemic.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
This work proposes a multimodal approach with which to predict the regional Gross Domestic Product (GDP) by combining historical GDP values with the embodied information in Twitter messages concerning the current economic condition. This proposal is of great interest, since it delivers forecasts at higher frequencies than both the official statistics (published only annually at the regional level in Spain) and the existing unofficial quarterly predictions (which rely on economic indicators that are available only after months of delay). The proposed method is based on a two-stage architecture. In the first stage, a multi-task autoencoder is initially used to obtain a GDP-related representation of tweets, which are then filtered to remove outliers and to obtain the GDP prediction from the consensus of opinions. In a second stage, this result is combined with the historical GDP values of the region using a multimodal network. The method is evaluated in four different regions of Spain using the tweets written by the most relevant economists, politicians, newspapers and institutions in each one. The results show that our approach successfully learns the evolution of the GDP using only historical information and tweets, thus making it possible to provide earlier forecasts about the regional GDP. This method also makes it possible to establish which the most or least influential opinions regarding this prediction are. As an additional exercise, we have assessed how well our method predicted the effect of the COVID-19 pandemic. López-Gutiérrez, J. C.; Valero-Mas, J. J.; Castellanos, F. J.; Calvo-Zaragoza, J.
Data Augmentation for End-to-End Optical Music Recognition Proceedings Article
In: Proceedings of the 14th IAPR International Workshop on Graphics Recognition (GREC), pp. 59-73, Springer, 2021.
BibTeX | Tags: GV/2020/030
@inproceedings{k473,
title = {Data Augmentation for End-to-End Optical Music Recognition},
author = {J. C. López-Gutiérrez and J. J. Valero-Mas and F. J. Castellanos and J. Calvo-Zaragoza},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
booktitle = {Proceedings of the 14th IAPR International Workshop on Graphics Recognition (GREC)},
pages = {59-73},
publisher = {Springer},
keywords = {GV/2020/030},
pubstate = {published},
tppubtype = {inproceedings}
}
Alfaro-Contreras, M.; Valero-Mas, J. J.; Iñesta, J. M.
Neural architectures for exploiting the components of Agnostic Notation in Optical Music Recognition Proceedings Article
In: Proc. of the 3rd International Workshop on Reading Music Systems, pp. 13-17, 2021.
BibTeX | Tags:
@inproceedings{k474,
title = {Neural architectures for exploiting the components of Agnostic Notation in Optical Music Recognition},
author = {M. Alfaro-Contreras and J. J. Valero-Mas and J. M. Iñesta},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
booktitle = {Proc. of the 3rd International Workshop on Reading Music Systems},
pages = {13-17},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Castellanos, F. J.; Gallego, A. J.; Calvo-Zaragoza, J.
Unsupervised Domain Adaptation for Document Analysis of Music Score Images Proceedings Article
In: Proc. of the 22nd International Society for Music Information Retrieval Conference, 2021.
@inproceedings{k475,
title = {Unsupervised Domain Adaptation for Document Analysis of Music Score Images},
author = {F. J. Castellanos and A. J. Gallego and J. Calvo-Zaragoza},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
booktitle = {Proc. of the 22nd International Society for Music Information Retrieval Conference},
keywords = {GRE19-04},
pubstate = {published},
tppubtype = {inproceedings}
}
Castellanos, F. J.; Valero-Mas, J. J.; Calvo-Zaragoza, J.
Prototype Generation in the String Space via Approximate Median for Data Reduction in Nearest Neighbor classification Journal Article
In: Soft Computing, vol. 25, 2021, ISSN: 15403-15415.
BibTeX | Tags: GRE19-04, HispaMus
@article{k476,
title = {Prototype Generation in the String Space via Approximate Median for Data Reduction in Nearest Neighbor classification},
author = {F. J. Castellanos and J. J. Valero-Mas and J. Calvo-Zaragoza},
issn = {15403-15415},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
journal = {Soft Computing},
volume = {25},
keywords = {GRE19-04, HispaMus},
pubstate = {published},
tppubtype = {article}
}
Ríos-Vila, A.; Rizo, D.; Calvo-Zaragoza, J.
Complete Optical Music Recognition via Agnostic Transcription and Machine Translation Proceedings Article
In: 16th International Conference on Document Analysis and Recognition, pp. 661-675, 2021.
BibTeX | Tags: GV/2020/030
@inproceedings{k477,
title = {Complete Optical Music Recognition via Agnostic Transcription and Machine Translation},
author = {A. Ríos-Vila and D. Rizo and J. Calvo-Zaragoza},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
booktitle = {16th International Conference on Document Analysis and Recognition},
pages = {661-675},
keywords = {GV/2020/030},
pubstate = {published},
tppubtype = {inproceedings}
}
Mas-Candela, E.; Alfaro-Contreras, M.; Calvo-Zaragoza, J.
Sequential Next-Symbol Prediction for Optical Music Recognition Proceedings Article
In: 16th International Conference on Document Analysis and Recognition, pp. 708-722, 2021, ISBN: 978-3-030-86334-0.
Links | BibTeX | Tags: GV/2020/030
@inproceedings{k478,
title = {Sequential Next-Symbol Prediction for Optical Music Recognition},
author = {E. Mas-Candela and M. Alfaro-Contreras and J. Calvo-Zaragoza},
url = {https://link.springer.com/chapter/10.1007/978-3-030-86334-0_46},
isbn = {978-3-030-86334-0},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
booktitle = {16th International Conference on Document Analysis and Recognition},
pages = {708-722},
keywords = {GV/2020/030},
pubstate = {published},
tppubtype = {inproceedings}
}
Garrido-Munoz, C.; Sánchez-Hernández, A.; Castellanos, F. J.; Calvo-Zaragoza, J.
Domain Adaptation for Document Image Binarization via Domain Classification Proceedings Article
In: Tallón-Ballesteros, A. J. (Ed.): Frontiers in Artificial Intelligence and Applications, pp. 569-582, IOS Press, 2021, ISBN: 978-1-64368-224-2.
BibTeX | Tags: GRE19-04, GV/2020/030
@inproceedings{k480,
title = {Domain Adaptation for Document Image Binarization via Domain Classification},
author = {C. Garrido-Munoz and A. Sánchez-Hernández and F. J. Castellanos and J. Calvo-Zaragoza},
editor = {A. J. Tallón-Ballesteros},
isbn = {978-1-64368-224-2},
year = {2021},
date = {2021-01-01},
booktitle = {Frontiers in Artificial Intelligence and Applications},
pages = {569-582},
publisher = {IOS Press},
chapter = {-},
keywords = {GRE19-04, GV/2020/030},
pubstate = {published},
tppubtype = {inproceedings}
}
Ortega-Bastida, J.
Aproximación de soluciones multimodales para aplicaciones Deep Learning PhD Thesis
2021.
BibTeX | Tags:
@phdthesis{k507,
title = {Aproximación de soluciones multimodales para aplicaciones Deep Learning},
author = {J. Ortega-Bastida},
editor = {J. Ramón Rico-Juan and A. J. Gallego},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
organization = {Universidad de Alicante},
keywords = {},
pubstate = {published},
tppubtype = {phdthesis}
}
Cuevas-Velasquez, H.; Gallego, A. J.; Fisher, R. B.
Two Heads are Better than One: Geometric-Latent Attention for Point Cloud Classification and Segmentation Proceedings Article
In: The 32nd British Machine Vision Conference (BMVC), 2021.
BibTeX | Tags:
@inproceedings{k513,
title = {Two Heads are Better than One: Geometric-Latent Attention for Point Cloud Classification and Segmentation},
author = {H. Cuevas-Velasquez and A. J. Gallego and R. B. Fisher},
year = {2021},
date = {2021-01-01},
booktitle = {The 32nd British Machine Vision Conference (BMVC)},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
2020
Bustos, A.; Pertusa, A.; Salinas, J. M.; Iglesia-Vayá, M.
PadChest: A large chest x-ray image dataset with multi-label annotated reports Journal Article
In: Medical Image Analysis, vol. 66, pp. 101797, 2020, ISSN: 1361-8415.
@article{k404,
title = {PadChest: A large chest x-ray image dataset with multi-label annotated reports},
author = {A. Bustos and A. Pertusa and J. M. Salinas and M. Iglesia-Vayá},
issn = {1361-8415},
year = {2020},
date = {2020-12-01},
journal = {Medical Image Analysis},
volume = {66},
pages = {101797},
abstract = {We present a labeled large-scale, high resolution chest x-ray dataset for the automated exploration of medical images along with their associated reports. This dataset includes more than 160,000 images obtained from 67,000 patients that were interpreted and reported by radiologists at San Juan Hospital (Spain) from 2009 to 2017, covering six different position views and additional information on image acquisition and patient demography. The reports were labeled with 174 different radiographic findings, 19 differential diagnoses and 104 anatomic locations organized as a hierarchical taxonomy and mapped onto standard Unified Medical Language System (UMLS) terminology. Of these reports, 27% were manually annotated by trained physicians and the remaining set was labeled using a supervised method based on a recurrent neural network with attention mechanisms. The labels generated were then validated in an independent test set achieving a 0.93 Micro-F1 score. To the best of our knowledge, this is one of the largest public chest x-ray databases suitable for training supervised models concerning radiographs, and the first to contain radiographic reports in Spanish. The PadChest dataset can be downloaded from http://bimcv.cipf.es/bimcv-projects/padchest/.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
We present a labeled large-scale, high resolution chest x-ray dataset for the automated exploration of medical images along with their associated reports. This dataset includes more than 160,000 images obtained from 67,000 patients that were interpreted and reported by radiologists at San Juan Hospital (Spain) from 2009 to 2017, covering six different position views and additional information on image acquisition and patient demography. The reports were labeled with 174 different radiographic findings, 19 differential diagnoses and 104 anatomic locations organized as a hierarchical taxonomy and mapped onto standard Unified Medical Language System (UMLS) terminology. Of these reports, 27% were manually annotated by trained physicians and the remaining set was labeled using a supervised method based on a recurrent neural network with attention mechanisms. The labels generated were then validated in an independent test set achieving a 0.93 Micro-F1 score. To the best of our knowledge, this is one of the largest public chest x-ray databases suitable for training supervised models concerning radiographs, and the first to contain radiographic reports in Spanish. The PadChest dataset can be downloaded from http://bimcv.cipf.es/bimcv-projects/padchest/. Rico-Juan, J. R.; Valero-Mas, J. J.; Iñesta, J. M.
Bounding Edit Distance for similarity-based sequence classification on Structural Pattern Recognition Journal Article
In: Applied Soft Computing, vol. 97 (Part A), 2020.
Abstract | BibTeX | Tags: HispaMus
@article{k451,
title = {Bounding Edit Distance for similarity-based sequence classification on Structural Pattern Recognition},
author = {J. R. Rico-Juan and J. J. Valero-Mas and J. M. Iñesta},
year = {2020},
date = {2020-12-01},
journal = {Applied Soft Computing},
volume = {97 (Part A)},
abstract = {Pattern Recognition tasks in the structural domain generally exhibit high accuracy results, but their time efficiency is quite low. Furthermore, this low performance is more pronounced when dealing with instance-based classifiers, since, for each query, the entire corpus must be evaluated to find the closest prototype. In this work we address this efficiency issue for the Nearest Neighbor classifier when data are encoded as two-dimensional code sequences, and more precisely strings and sequences of vectors. For this, a set of bounds is proposed in the distance metric that avoid the calculation of unnecessary distances. Results obtained prove the effectiveness of the proposal as it reduces the classification time in percentages between 80% and 90% for string representations and between 60% and 80% for data codified as sequences of vectors with respect to their corresponding non-optimized version of the classifier.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {article}
}
Pattern Recognition tasks in the structural domain generally exhibit high accuracy results, but their time efficiency is quite low. Furthermore, this low performance is more pronounced when dealing with instance-based classifiers, since, for each query, the entire corpus must be evaluated to find the closest prototype. In this work we address this efficiency issue for the Nearest Neighbor classifier when data are encoded as two-dimensional code sequences, and more precisely strings and sequences of vectors. For this, a set of bounds is proposed in the distance metric that avoid the calculation of unnecessary distances. Results obtained prove the effectiveness of the proposal as it reduces the classification time in percentages between 80% and 90% for string representations and between 60% and 80% for data codified as sequences of vectors with respect to their corresponding non-optimized version of the classifier. Saborit-Torres, J. M.; Saenz-Gamboa, J. J.; Montell, J. À.; Salinas, J. M.; Gómez, J. A.; Stefan, I.; Caparrós, M.; García-García, F.; Domenech, J.; Manjón, J. V.; Rojas, G.; Pertusa, A.; Bustos, A.; González, G.; Galant, J.; Iglesia-Vayá, M.
Medical imaging data structure extended to multiple modalities and anatomical regions Journal Article
In: ArXiv:2010.00434, 2020.
@article{k452,
title = {Medical imaging data structure extended to multiple modalities and anatomical regions},
author = {J. M. Saborit-Torres and J. J. Saenz-Gamboa and J. À. Montell and J. M. Salinas and J. A. Gómez and I. Stefan and M. Caparrós and F. García-García and J. Domenech and J. V. Manjón and G. Rojas and A. Pertusa and A. Bustos and G. González and J. Galant and M. Iglesia-Vayá},
year = {2020},
date = {2020-10-01},
journal = {ArXiv:2010.00434},
abstract = {Brain Imaging Data Structure (BIDS) allows the user to organise brain imaging data into a clear and easy standard directory structure. BIDS is widely supported by the scientific community and is considered a powerful standard for management. The original BIDS is limited to images or data related to the brain. Medical Imaging Data Structure (MIDS) was therefore conceived with the objective of extending this methodology to other anatomical regions and other types of imaging systems in these areas.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Brain Imaging Data Structure (BIDS) allows the user to organise brain imaging data into a clear and easy standard directory structure. BIDS is widely supported by the scientific community and is considered a powerful standard for management. The original BIDS is limited to images or data related to the brain. Medical Imaging Data Structure (MIDS) was therefore conceived with the objective of extending this methodology to other anatomical regions and other types of imaging systems in these areas. Ríos-Vila, A.; Calvo-Zaragoza, J.; Rizo, D.
Evaluating Simultaneous Recognition and Encoding for Optical Music Recognition Proceedings Article
In: DLfM 2020: 7th International Conference on Digital Libraries for Musicology, pp. 10-17, Association for Computing Machinery, 2020, ISBN: 978-1-4503-8760-6.
@inproceedings{k456,
title = {Evaluating Simultaneous Recognition and Encoding for Optical Music Recognition},
author = {A. Ríos-Vila and J. Calvo-Zaragoza and D. Rizo},
isbn = {978-1-4503-8760-6},
year = {2020},
date = {2020-10-01},
booktitle = {DLfM 2020: 7th International Conference on Digital Libraries for Musicology},
journal = {DLfM 2020: 7th International Conference on Digital Libraries for Musicology},
pages = {10-17},
publisher = {Association for Computing Machinery},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Castellanos, F. J.; Calvo-Zaragoza, J.; Iñesta, J. M.
A neural approach for full-page Optical Music Recognition of mensural documents Proceedings Article
In: Proceedings of the 21st International Society for Music Information Retrieval Conference, pp. 558-565, Montréal, Canada, 2020, ISBN: 978-0-9813537-0-8.
@inproceedings{k457,
title = {A neural approach for full-page Optical Music Recognition of mensural documents},
author = {F. J. Castellanos and J. Calvo-Zaragoza and J. M. Iñesta},
isbn = {978-0-9813537-0-8},
year = {2020},
date = {2020-10-01},
booktitle = {Proceedings of the 21st International Society for Music Information Retrieval Conference},
pages = {558-565},
address = {Montréal, Canada},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Castellanos, F. J.; Calvo-Zaragoza, J.; Iñesta, J. M.
Two-step neural cross-domain experiments forfull-page recognition of Mensural documents Proceedings Article
In: Proc. of the 13th International Workshop on Machine Learning and Music, pp. 56–60, MML 2020 13th International Workshop on Machine Learning and Music, 2020.
BibTeX | Tags: GRE19-04, HispaMus
@inproceedings{k448,
title = {Two-step neural cross-domain experiments forfull-page recognition of Mensural documents},
author = {F. J. Castellanos and J. Calvo-Zaragoza and J. M. Iñesta},
year = {2020},
date = {2020-09-01},
booktitle = {Proc. of the 13th International Workshop on Machine Learning and Music},
pages = {56--60},
publisher = {13th International Workshop on Machine Learning and Music},
organization = {MML 2020},
keywords = {GRE19-04, HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Castellanos, F. J.; Gallego, A. J.; Calvo-Zaragoza, J.
Automatic scale estimation for music score images Journal Article
In: Expert Systems with Applications, vol. 158, no. 113590, 2020, ISSN: 0957-4174.
BibTeX | Tags:
@article{k450,
title = {Automatic scale estimation for music score images},
author = {F. J. Castellanos and A. J. Gallego and J. Calvo-Zaragoza},
issn = {0957-4174},
year = {2020},
date = {2020-06-01},
urldate = {2020-06-01},
journal = {Expert Systems with Applications},
volume = {158},
number = {113590},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Vayá, Saborit M. Iglesia; Montell, J. A.; Pertusa, A.; Bustos, A.; Cazorla, M.; Galant, J.; Barber, X.; Orozco-Beltrán, D.; García-García, F.; Caparrós, M.; González, G.; Salinas, J. M.
BIMCV COVID-19+: a large annotated dataset of RX and CT images from COVID-19 patients Journal Article
In: ArXiv:2006.01174, 2020.
@article{k453,
title = {BIMCV COVID-19+: a large annotated dataset of RX and CT images from COVID-19 patients},
author = {Saborit M. Iglesia Vayá and J. A. Montell and A. Pertusa and A. Bustos and M. Cazorla and J. Galant and X. Barber and D. Orozco-Beltrán and F. García-García and M. Caparrós and G. González and J. M. Salinas},
year = {2020},
date = {2020-06-01},
journal = {ArXiv:2006.01174},
abstract = {This paper describes BIMCV COVID-19+, a large dataset from the Valencian Region Medical ImageBank (BIMCV) containing chest X-ray images CXR (CR, DX) and computed tomography (CT) imaging of COVID-19+ patients along with their radiological findings and locations, pathologies, radiological reports (in Spanish), DICOM metadata, Polymerase chain reaction (PCR), Immunoglobulin G (IgG) and Immunoglobulin M (IgM) diagnostic antibody tests. The findings have been mapped onto standard Unified Medical Language System (UMLS) terminology and cover a wide spectrum of thoracic entities, unlike the considerably more reduced number of entities annotated in previous datasets. Images are stored in high resolution and entities are localized with anatomical labels and stored in a Medical Imaging Data Structure (MIDS) format. In addition, 10 images were annotated by a team of radiologists to include semantic segmentation of radiological findings. This first iteration of the database includes 1,380 CX, 885 DX and 163 CT studies from 1,311 COVID-19+ patients. This is, to the best of our knowledge, the largest COVID-19+ dataset of images available in an open format. The dataset can be downloaded from this http URL.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
This paper describes BIMCV COVID-19+, a large dataset from the Valencian Region Medical ImageBank (BIMCV) containing chest X-ray images CXR (CR, DX) and computed tomography (CT) imaging of COVID-19+ patients along with their radiological findings and locations, pathologies, radiological reports (in Spanish), DICOM metadata, Polymerase chain reaction (PCR), Immunoglobulin G (IgG) and Immunoglobulin M (IgM) diagnostic antibody tests. The findings have been mapped onto standard Unified Medical Language System (UMLS) terminology and cover a wide spectrum of thoracic entities, unlike the considerably more reduced number of entities annotated in previous datasets. Images are stored in high resolution and entities are localized with anatomical labels and stored in a Medical Imaging Data Structure (MIDS) format. In addition, 10 images were annotated by a team of radiologists to include semantic segmentation of radiological findings. This first iteration of the database includes 1,380 CX, 885 DX and 163 CT studies from 1,311 COVID-19+ patients. This is, to the best of our knowledge, the largest COVID-19+ dataset of images available in an open format. The dataset can be downloaded from this http URL. González, G.; Bustos, A.; Salinas, J. M.; Iglesia-Vaya, M.; Galant, J.; Cano-Espinosa, C.; Barber, X.; Orozco-Beltrán, D.; Cazorla, M.; Pertusa, A.
UMLS-ChestNet: A deep convolutional neural network for radiological findings, differential diagnoses and localizations of COVID-19 in chest x-rays Journal Article
In: arxiv:2006.05274, 2020.
@article{k454,
title = {UMLS-ChestNet: A deep convolutional neural network for radiological findings, differential diagnoses and localizations of COVID-19 in chest x-rays},
author = {G. González and A. Bustos and J. M. Salinas and M. Iglesia-Vaya and J. Galant and C. Cano-Espinosa and X. Barber and D. Orozco-Beltrán and M. Cazorla and A. Pertusa},
year = {2020},
date = {2020-06-01},
journal = {arxiv:2006.05274},
abstract = {In this work we present a method for the detection of radiological findings, their location and differential diagnoses from chest x-rays. Unlike prior works that focus on the detection of few pathologies, we use a hierarchical taxonomy mapped to the Unified Medical Language System (UMLS) terminology to identify 189 radiological findings, 22 differential diagnosis and 122 anatomic locations, including ground glass opacities, infiltrates, consolidations and other radiological findings compatible with COVID-19. We train the system on one large database of 92,594 frontal chest x-rays (AP or PA, standing, supine or decubitus) and a second database of 2,065 frontal images of COVID-19 patients identified by at least one positive Polymerase Chain Reaction (PCR) test. The reference labels are obtained through natural language processing of the radiological reports. On 23,159 test images, the proposed neural network obtains an AUC of 0.94 for the diagnosis of COVID-19. To our knowledge, this work uses the largest chest x-ray dataset of COVID-19 positive cases to date and is the first one to use a hierarchical labeling schema and to provide interpretability of the results, not only by using network attention methods, but also by indicating the radiological findings that have led to the diagnosis.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
In this work we present a method for the detection of radiological findings, their location and differential diagnoses from chest x-rays. Unlike prior works that focus on the detection of few pathologies, we use a hierarchical taxonomy mapped to the Unified Medical Language System (UMLS) terminology to identify 189 radiological findings, 22 differential diagnosis and 122 anatomic locations, including ground glass opacities, infiltrates, consolidations and other radiological findings compatible with COVID-19. We train the system on one large database of 92,594 frontal chest x-rays (AP or PA, standing, supine or decubitus) and a second database of 2,065 frontal images of COVID-19 patients identified by at least one positive Polymerase Chain Reaction (PCR) test. The reference labels are obtained through natural language processing of the radiological reports. On 23,159 test images, the proposed neural network obtains an AUC of 0.94 for the diagnosis of COVID-19. To our knowledge, this work uses the largest chest x-ray dataset of COVID-19 positive cases to date and is the first one to use a hierarchical labeling schema and to provide interpretability of the results, not only by using network attention methods, but also by indicating the radiological findings that have led to the diagnosis. Thomae, M. E.; Ríos-Vila, A.; Calvo-Zaragoza, J.; Rizo, D.; Iñesta, J. M.
Retrieving Music Semantics from Optical Music Recognition by Machine Translation Proceedings Article
In: Luca, Elsa De; Flanders, Julia (Ed.): Music Encoding Conference Proceedings 26-29 May, 2020 Tufts University, Boston (USA), pp. 19-24, Music Encoding Initiative 2020.
Links | BibTeX | Tags: HispaMus
@inproceedings{k446,
title = {Retrieving Music Semantics from Optical Music Recognition by Machine Translation},
author = {M. E. Thomae and A. Ríos-Vila and J. Calvo-Zaragoza and D. Rizo and J. M. Iñesta},
editor = {Elsa De Luca and Julia Flanders},
url = {https://grfia.dlsi.ua.es/repositori/grfia/pubs/446/music_encoding_conference_proceedings_2020.pdf},
year = {2020},
date = {2020-05-01},
booktitle = {Music Encoding Conference Proceedings 26-29 May, 2020 Tufts University, Boston (USA)},
pages = {19-24},
organization = {Music Encoding Initiative},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Martinez-Martin, E.; Pertusa, A.; Costa, A.; Gomez-Donoso, F.; Escalona, F.; Viejo, D.; Cazorla, M.
The competition as assessment tool in robotics engineering Proceedings Article
In: INTED2020 Proceedings, pp. 1082-1085, Valencia, Spain, 2020, ISBN: 978-84-09-17939-8.
BibTeX | Tags:
@inproceedings{k460,
title = {The competition as assessment tool in robotics engineering},
author = {E. Martinez-Martin and A. Pertusa and A. Costa and F. Gomez-Donoso and F. Escalona and D. Viejo and M. Cazorla},
isbn = {978-84-09-17939-8},
year = {2020},
date = {2020-03-01},
booktitle = {INTED2020 Proceedings},
pages = {1082-1085},
address = {Valencia, Spain},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Calvo-Zaragoza, J.; Jr., J. Hajic; Pacha, A.
Understanding Optical Music Recognition Journal Article
In: ACM Computing Surveys, vol. 53, no. 4, pp. 77, 2020.
@article{k426,
title = {Understanding Optical Music Recognition},
author = {J. Calvo-Zaragoza and J. Hajic Jr. and A. Pacha},
year = {2020},
date = {2020-01-01},
journal = {ACM Computing Surveys},
volume = {53},
number = {4},
pages = {77},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {article}
}
Cuevas-Velasquez, H.; Gallego, A. J.; Fisher, R. B.
Segmentation and 3D reconstruction of rose plants from stereoscopic images Journal Article
In: Computers and Electronics in Agriculture, vol. 171, pp. 105296, 2020, ISSN: 0168-1699.
@article{k440,
title = {Segmentation and 3D reconstruction of rose plants from stereoscopic images},
author = {H. Cuevas-Velasquez and A. J. Gallego and R. B. Fisher},
issn = {0168-1699},
year = {2020},
date = {2020-01-01},
urldate = {2020-01-01},
journal = {Computers and Electronics in Agriculture},
volume = {171},
pages = {105296},
abstract = {The method proposed in this paper is part of the vision module of a garden robot capable of navigating towards rose bushes and clip them according to a set of pruning rules. The method is responsible for performing the segmentation of the branches and recovering their morphology in 3D. The obtained reconstruction allows the manipulator of the robot to select the candidate branches to be pruned. This method first obtains a stereo pair of images and calculates the disparity image using block matching and the segmentation of the branches using a Fully Convolutional Neuronal Network modified to return a map with the probability at the pixel level of the presence of a branch. A post-processing step combines the segmentation and the disparity in order to improve the results. Then, the skeleton of the plant and the branching structure are calculated, and finally, the 3D reconstruction is obtained. The proposed approach is evaluated with five different datasets, three of them compiled by the authors and two from the state of the art, including indoor and outdoor scenes with uncontrolled environments. The different steps of the proposed pipeline are evaluated and compared with other state-of-the-art methods, showing that the accuracy of the segmentation improves other methods for this task, even with variable lighting, and also that the skeletonization and the reconstruction processes obtain robust results.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The method proposed in this paper is part of the vision module of a garden robot capable of navigating towards rose bushes and clip them according to a set of pruning rules. The method is responsible for performing the segmentation of the branches and recovering their morphology in 3D. The obtained reconstruction allows the manipulator of the robot to select the candidate branches to be pruned. This method first obtains a stereo pair of images and calculates the disparity image using block matching and the segmentation of the branches using a Fully Convolutional Neuronal Network modified to return a map with the probability at the pixel level of the presence of a branch. A post-processing step combines the segmentation and the disparity in order to improve the results. Then, the skeleton of the plant and the branching structure are calculated, and finally, the 3D reconstruction is obtained. The proposed approach is evaluated with five different datasets, three of them compiled by the authors and two from the state of the art, including indoor and outdoor scenes with uncontrolled environments. The different steps of the proposed pipeline are evaluated and compared with other state-of-the-art methods, showing that the accuracy of the segmentation improves other methods for this task, even with variable lighting, and also that the skeletonization and the reconstruction processes obtain robust results. Cuevas-Velasquez, H.; Gallego, A. J.; Tylecek, R.; Hemming, J.; Tuijl, B.; Mencarelli, A.; Fisher, R. B.
Real-time Stereo Visual Servoing for Rose Pruning with Robotic Arm Proceedings Article
In: International Conference on Robotics and Automation (ICRA), IEEE, Paris, France, 2020.
@inproceedings{k441,
title = {Real-time Stereo Visual Servoing for Rose Pruning with Robotic Arm},
author = {H. Cuevas-Velasquez and A. J. Gallego and R. Tylecek and J. Hemming and B. Tuijl and A. Mencarelli and R. B. Fisher},
year = {2020},
date = {2020-01-01},
urldate = {2020-01-01},
booktitle = {International Conference on Robotics and Automation (ICRA)},
publisher = {IEEE},
address = {Paris, France},
abstract = {The paper presents a working pipeline which integrates hardware and software in an automated robotic rose cutter. To the best of our knowledge, this is the first robot able to prune rose bushes in a natural environment. Unlike similar approaches like tree stem cutting, the proposed method does not require to scan the full plant, have multiple cameras around the bush, or assume that a stem does not move. It relies on a single stereo camera mounted on the end-effector of the robot and real-time visual servoing to navigate to the desired cutting location on the stem. The evaluation of the whole pipeline shows a good performance in a garden with unconstrained conditions, where finding and approaching a specific location on a stem is challenging due to occlusions caused by other stems and dynamic changes caused by the wind.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The paper presents a working pipeline which integrates hardware and software in an automated robotic rose cutter. To the best of our knowledge, this is the first robot able to prune rose bushes in a natural environment. Unlike similar approaches like tree stem cutting, the proposed method does not require to scan the full plant, have multiple cameras around the bush, or assume that a stem does not move. It relies on a single stereo camera mounted on the end-effector of the robot and real-time visual servoing to navigate to the desired cutting location on the stem. The evaluation of the whole pipeline shows a good performance in a garden with unconstrained conditions, where finding and approaching a specific location on a stem is challenging due to occlusions caused by other stems and dynamic changes caused by the wind. Oncina, J.; Iñesta, J. M.; Micó, L.
Adaptively Learning to Recognize Symbols in Handwritten Early Music Book Chapter
In: Cellier, Driessens Peggy (Ed.): Machine Learning and Knowledge Discovery in Databases, Chapter 40, pp. 470–477, Springer, 2020, ISBN: 978-3-030-43887-6.
Abstract | BibTeX | Tags: HispaMus
@inbook{k443,
title = {Adaptively Learning to Recognize Symbols in Handwritten Early Music},
author = {J. Oncina and J. M. Iñesta and L. Micó},
editor = {Driessens Peggy Cellier},
isbn = {978-3-030-43887-6},
year = {2020},
date = {2020-01-01},
urldate = {2020-01-01},
booktitle = {Machine Learning and Knowledge Discovery in Databases},
pages = {470--477},
publisher = {Springer},
chapter = {40},
abstract = {Human supervision is necessary for a correct edition and publication of handwritten early music collections. The output of an optical music recognition system for that kind of documents may contain a significant number of errors, making it tedious to correct for a human expert. An adequate strategy is needed to optimize the human feedback information during the correction stage to adapt the classifier to the specificities of each manuscript. In this paper, we compare the performance of a neural system, difficult and slow to be retrained, and a nearest neighbor strategy, based on the neural codes provided by a neural net, trained offline, used as a feature extractor.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inbook}
}
Human supervision is necessary for a correct edition and publication of handwritten early music collections. The output of an optical music recognition system for that kind of documents may contain a significant number of errors, making it tedious to correct for a human expert. An adequate strategy is needed to optimize the human feedback information during the correction stage to adapt the classifier to the specificities of each manuscript. In this paper, we compare the performance of a neural system, difficult and slow to be retrained, and a nearest neighbor strategy, based on the neural codes provided by a neural net, trained offline, used as a feature extractor. Gallego, A. J.; Calvo-Zaragoza, J.; Rico-Juan, J. R.
Insights Into Efficient k-Nearest Neighbor Classification With Convolutional Neural Codes Journal Article
In: IEEE Access, vol. 8, pp. 99312-99326, 2020, ISSN: 2169-3536.
Abstract | BibTeX | Tags: HispaMus
@article{k444,
title = {Insights Into Efficient k-Nearest Neighbor Classification With Convolutional Neural Codes},
author = {A. J. Gallego and J. Calvo-Zaragoza and J. R. Rico-Juan},
issn = {2169-3536},
year = {2020},
date = {2020-01-01},
urldate = {2020-01-01},
journal = {IEEE Access},
volume = {8},
pages = {99312-99326},
abstract = {The increasing consideration of Convolutional Neural Networks (CNN) has not prevented the use of the k-Nearest Neighbor (kNN) method. In fact, a hybrid CNN-kNN approach is an interesting option in which the network specializes in feature extraction through its activations (Neural Codes), while the kNN has the advantage of performing a retrieval by means of similarity. However, this hybrid approach also has the disadvantages of the kNN search, and especially its high computational cost which is, in principle, undesirable for large-scale data. In this paper, we present the first comprehensive study of efficient kNN search algorithms using this hybrid CNN-kNN approach. This has been done by considering up to 16 different algorithms, each of which is evaluated with a different parametrization, in 7 datasets of heterogeneous composition. Our results show that no single algorithm is capable of covering all aspects, but rather that each family of algorithms is better suited to specific aspects of the problem. This signifies that Fast Similarity Search algorithms maintain their performance, but do not reduce the cost as much as the Data Reduction family does. In turn, the Approximated Similarity Search family is postulated as a good option when attempting to balance accuracy and efficiency. The experiments also suggest that considering statistical transformation algorithms such as Linear Discriminant Analysis might be useful in certain cases.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {article}
}
The increasing consideration of Convolutional Neural Networks (CNN) has not prevented the use of the k-Nearest Neighbor (kNN) method. In fact, a hybrid CNN-kNN approach is an interesting option in which the network specializes in feature extraction through its activations (Neural Codes), while the kNN has the advantage of performing a retrieval by means of similarity. However, this hybrid approach also has the disadvantages of the kNN search, and especially its high computational cost which is, in principle, undesirable for large-scale data. In this paper, we present the first comprehensive study of efficient kNN search algorithms using this hybrid CNN-kNN approach. This has been done by considering up to 16 different algorithms, each of which is evaluated with a different parametrization, in 7 datasets of heterogeneous composition. Our results show that no single algorithm is capable of covering all aspects, but rather that each family of algorithms is better suited to specific aspects of the problem. This signifies that Fast Similarity Search algorithms maintain their performance, but do not reduce the cost as much as the Data Reduction family does. In turn, the Approximated Similarity Search family is postulated as a good option when attempting to balance accuracy and efficiency. The experiments also suggest that considering statistical transformation algorithms such as Linear Discriminant Analysis might be useful in certain cases. Valero-Mas, J. J.; Castellanos, F. J.
Data Reduction in the String Space for Efficient kNN Classification through Space Partitioning Journal Article
In: Applied Sciences, vol. 10, no. 10, pp. 3356, 2020.
@article{k445,
title = {Data Reduction in the String Space for Efficient kNN Classification through Space Partitioning},
author = {J. J. Valero-Mas and F. J. Castellanos},
year = {2020},
date = {2020-01-01},
journal = {Applied Sciences},
volume = {10},
number = {10},
pages = {3356},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {article}
}
Pertusa, A.; Román, M. A.
Data representations for audio-to-score monophonic music transcription Journal Article
In: Expert Systems with Applications, vol. 162, pp. 113769, 2020, ISSN: 0957-4174.
@article{k447,
title = {Data representations for audio-to-score monophonic music transcription},
author = {A. Pertusa and M. A. Román},
issn = {0957-4174},
year = {2020},
date = {2020-01-01},
urldate = {2020-01-01},
journal = {Expert Systems with Applications},
volume = {162},
pages = {113769},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {article}
}
Ríos-Vila, A.; Calvo-Zaragoza, J.; Iñesta, J. M.
Exploring the two-dimensional nature of music notation for score recognition with end-to-end approaches Proceedings Article
In: Proc. of 17th International Conference on Frontiers in Handwriting Recognition (ICFHR), pp. 193–198, IEEE Computer Society IEEE, Dormund (Germany), 2020, ISBN: 978-1-7281-9966-5.
Abstract | BibTeX | Tags: HispaMus
@inproceedings{k461,
title = {Exploring the two-dimensional nature of music notation for score recognition with end-to-end approaches},
author = {A. Ríos-Vila and J. Calvo-Zaragoza and J. M. Iñesta},
isbn = {978-1-7281-9966-5},
year = {2020},
date = {2020-01-01},
booktitle = {Proc. of 17th International Conference on Frontiers in Handwriting Recognition (ICFHR)},
pages = {193--198},
publisher = {IEEE},
address = {Dormund (Germany)},
organization = {IEEE Computer Society},
abstract = {Optical Music Recognition workflows perform several steps to retrieve the content in music score images, being symbol recognition one of the key stages. State-of-the-art approaches for this stage currently address the coding of the output symbols as if they were plain text characters. However, music symbols have a two-dimensional nature that is ignored in these approaches. In this paper, we explore alternative output representations to perform music symbol recognition with state-of-the-art end-to-end neural technologies. We propose and describe new output representations which take into account the mentioned two-dimensional nature. We seek answers to the question of whether it is possible to obtain better recognition results in both printed and handwritten music scores. In this analysis, we compare the results given using three output encodings and two neural approaches. We found that one of the proposed encodings outperforms the results obtained by the standard one. This permits us to conclude that it is interesting to keep researching on this topic to improve end-to-end music score recognition.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Optical Music Recognition workflows perform several steps to retrieve the content in music score images, being symbol recognition one of the key stages. State-of-the-art approaches for this stage currently address the coding of the output symbols as if they were plain text characters. However, music symbols have a two-dimensional nature that is ignored in these approaches. In this paper, we explore alternative output representations to perform music symbol recognition with state-of-the-art end-to-end neural technologies. We propose and describe new output representations which take into account the mentioned two-dimensional nature. We seek answers to the question of whether it is possible to obtain better recognition results in both printed and handwritten music scores. In this analysis, we compare the results given using three output encodings and two neural approaches. We found that one of the proposed encodings outperforms the results obtained by the standard one. This permits us to conclude that it is interesting to keep researching on this topic to improve end-to-end music score recognition. Rizo, D.; Calvo-Zaragoza, J.; Iñesta, J. M.
Preservación del patrimonio musical mediante transcripción asistida por ordenador Book Chapter
In: Isusi-Fagoaga, R.; Serrano, F. Villanueva (Ed.): La música de la Corona d’Aragó: investigació, transferència i educació, Chapter 15, pp. 275–298, Universitat de València, 2020, ISBN: 978-84-09-19985-3.
@inbook{k466,
title = {Preservación del patrimonio musical mediante transcripción asistida por ordenador},
author = {D. Rizo and J. Calvo-Zaragoza and J. M. Iñesta},
editor = {R. Isusi-Fagoaga and F. Villanueva Serrano},
isbn = {978-84-09-19985-3},
year = {2020},
date = {2020-01-01},
urldate = {2020-01-01},
booktitle = {La música de la Corona d’Aragó: investigació, transferència i educació},
pages = {275--298},
publisher = {Universitat de València},
chapter = {15},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inbook}
}
Ortega-Bastida, J.; Gallego, A. J.; Rico-Juan, J. R.; Albarrán, P.
Regional gross domestic product prediction using Twitter deep learning representations Proceedings Article
In: 17th IADIS International Conference on Applied Computing (IADIS AC), pp. 89–96, Lisboa, Portugal, 2020, ISBN: 978-989-8704-24-5.
BibTeX | Tags:
@inproceedings{k472,
title = {Regional gross domestic product prediction using Twitter deep learning representations},
author = {J. Ortega-Bastida and A. J. Gallego and J. R. Rico-Juan and P. Albarrán},
isbn = {978-989-8704-24-5},
year = {2020},
date = {2020-01-01},
urldate = {2020-01-01},
booktitle = {17th IADIS International Conference on Applied Computing (IADIS AC)},
pages = {89--96},
address = {Lisboa, Portugal},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
2019
Calvo-Zaragoza, J.; Toselli, A. H.; Vidal, E.
Handwritten Music Recognition for Mensural Notation with Convolutional Recurrent Neural Networks Journal Article
In: Pattern Recognition Letters, vol. 128, pp. 115–121, 2019.
Links | BibTeX | Tags: HispaMus
@article{k415,
title = {Handwritten Music Recognition for Mensural Notation with Convolutional Recurrent Neural Networks},
author = {J. Calvo-Zaragoza and A. H. Toselli and E. Vidal},
url = {https://grfia.dlsi.ua.es/repositori/grfia/pubs/415/1-s2.0-S0167865519302338-main.pdf},
year = {2019},
date = {2019-12-01},
journal = {Pattern Recognition Letters},
volume = {128},
pages = {115--121},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {article}
}
Iñesta, J. M.; Rizo, D.; Calvo-Zaragoza, J.
MuRET as a software for the transcription of historical archives Proceedings Article
In: Proceedings of the 2nd Workshop on Reading Music Systems, WoRMS, pp. 12–15, Delft (The Nederlands), 2019.
Abstract | Links | BibTeX | Tags: HispaMus
@inproceedings{k437,
title = {MuRET as a software for the transcription of historical archives},
author = {J. M. Iñesta and D. Rizo and J. Calvo-Zaragoza},
url = {https://grfia.dlsi.ua.es/repositori/grfia/pubs/437/worms-2019-inesta.pdf},
year = {2019},
date = {2019-11-01},
booktitle = {Proceedings of the 2nd Workshop on Reading Music Systems, WoRMS},
pages = {12--15},
address = {Delft (The Nederlands)},
abstract = {The transcription process from historical hand- written music manuscripts to a structured digital encoding has been traditionally performed following a fully manual workflow. At most, it has received some technological support in particular stages, like optical music recognition (OMR) of the source images, or transcription to modern notation with music edition applications. Currently, there is no mature and stable enough solution for the OMR problem, and the most used music editors do not support early notations, such as the mensural notation. A new tool called MUsic Recognition, Encoding, and Transcription (MuRET) has been developed, which covers all transcription phases, from the manuscript image to the encoded digital content. MuRET is designed as a machine-learning based research tool, allowing different processing approaches to be used, and producing both the expected transcribed contents in standard encodings and data for the study of the transcription process.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
The transcription process from historical hand- written music manuscripts to a structured digital encoding has been traditionally performed following a fully manual workflow. At most, it has received some technological support in particular stages, like optical music recognition (OMR) of the source images, or transcription to modern notation with music edition applications. Currently, there is no mature and stable enough solution for the OMR problem, and the most used music editors do not support early notations, such as the mensural notation. A new tool called MUsic Recognition, Encoding, and Transcription (MuRET) has been developed, which covers all transcription phases, from the manuscript image to the encoded digital content. MuRET is designed as a machine-learning based research tool, allowing different processing approaches to be used, and producing both the expected transcribed contents in standard encodings and data for the study of the transcription process. Rizo, D.
DLfM '19: 6th International Conference on Digital Libraries for Musicology Book
ACM, The Hague, Netherlands, 2019, ISBN: 978-1-4503-7239-8.
@book{k438,
title = {DLfM '19: 6th International Conference on Digital Libraries for Musicology},
author = {D. Rizo},
editor = {D. Rizo},
isbn = {978-1-4503-7239-8},
year = {2019},
date = {2019-11-01},
urldate = {2019-11-01},
publisher = {ACM},
address = {The Hague, Netherlands},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {book}
}
Ríos-Vila, A.; Calvo-Zaragoza, J.; Rizo, D.; Iñesta, J. M.
ReadSco: An Open-Source Web-Based Optical Music Recognition Tool Proceedings Article
In: Calvo-Zaragoza, J.; Pacha, A. (Ed.): Proc. of the 2nd International Workshop on Reading Music Systems, WoRMS 2019, pp. 27-30, Delft (The Netherlands), 2019.
Abstract | Links | BibTeX | Tags: HispaMus
@inproceedings{k439,
title = {ReadSco: An Open-Source Web-Based Optical Music Recognition Tool},
author = {A. Ríos-Vila and J. Calvo-Zaragoza and D. Rizo and J. M. Iñesta},
editor = {J. Calvo-Zaragoza and A. Pacha},
url = {https://grfia.dlsi.ua.es/repositori/grfia/pubs/439/ReadSco_Worms2019.pdf},
year = {2019},
date = {2019-11-01},
urldate = {2019-11-01},
booktitle = {Proc. of the 2nd International Workshop on Reading Music Systems, WoRMS 2019},
pages = {27-30},
address = {Delft (The Netherlands)},
abstract = {We introduce READSCO, an open-source web-based community tool for Optical Music Recognition. READSCO aims to serve as a connection between research results and practical use. We describe the design decisions considered to both favours a rapid integration of new advances from the research field and
facilitate the community’s participation in its development. The project is still in its planning phase, so this work is a good opportunity to present the main idea and get direct feedback from other researchers.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
We introduce READSCO, an open-source web-based community tool for Optical Music Recognition. READSCO aims to serve as a connection between research results and practical use. We describe the design decisions considered to both favours a rapid integration of new advances from the research field and
facilitate the community’s participation in its development. The project is still in its planning phase, so this work is a good opportunity to present the main idea and get direct feedback from other researchers. Iñesta, J. M.; Rizo, D.; Calvo-Zaragoza, J.
Recognition, encoding, and transcription of early mensural handwritten music Proceedings Article
In: procedings of the 13th International Workshop on Graphics Recognition (GREC 2019), pp. 13-14, Sydney (Australia), 2019.
@inproceedings{k416,
title = {Recognition, encoding, and transcription of early mensural handwritten music},
author = {J. M. Iñesta and D. Rizo and J. Calvo-Zaragoza},
year = {2019},
date = {2019-09-01},
booktitle = {procedings of the 13th International Workshop on Graphics Recognition (GREC 2019)},
pages = {13-14},
address = {Sydney (Australia)},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Micó, L.; Oncina, J.; Iñesta, J. M.
Adaptively learning to recognize symbols in handwritten early music Book Chapter
In: Conklin, D. (Ed.): procedings of the ECML/PKDD 12th International Workshop on Machine Learning and Music (MML 2019), Chapter (to appear), Springer, Würzburg (Germany), 2019.
Abstract | Links | BibTeX | Tags: HispaMus
@inbook{k417,
title = {Adaptively learning to recognize symbols in handwritten early music},
author = {L. Micó and J. Oncina and J. M. Iñesta},
editor = {D. Conklin},
url = {https://grfia.dlsi.ua.es/repositori/grfia/pubs/417/Learning_to_recognize_handwritten_early_music.pdf},
year = {2019},
date = {2019-09-01},
booktitle = {procedings of the ECML/PKDD 12th International Workshop on Machine Learning and Music (MML 2019)},
publisher = {Springer},
address = {Würzburg (Germany)},
chapter = {(to appear)},
abstract = {Human supervision is necessary for a correct edition and publication of handwritten early music collections. The output of an op- tical music recognition system for that kind of documents may contain a significant number of errors, making it tedious to correct for a human expert. An adequate strategy is needed to optimize the human feedback information during the correction stage to adapt the classifier to the specificities of each manuscript. In this paper, we compare the perfor- mance of a neural system, difficult and slow to be retrained, and a nearest neighbor strategy, based on the neural codes provided by a neural net, trained offline, used as a feature extractor.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inbook}
}
Human supervision is necessary for a correct edition and publication of handwritten early music collections. The output of an op- tical music recognition system for that kind of documents may contain a significant number of errors, making it tedious to correct for a human expert. An adequate strategy is needed to optimize the human feedback information during the correction stage to adapt the classifier to the specificities of each manuscript. In this paper, we compare the perfor- mance of a neural system, difficult and slow to be retrained, and a nearest neighbor strategy, based on the neural codes provided by a neural net, trained offline, used as a feature extractor. Alfaro-Contreras, M.; Calvo-Zaragoza, J.; Iñesta, J. M.
Approaching End-to-end Optical Music Recognition for Homophonic Scores Proceedings Article
In: Iberian Conference on Pattern Recognition and Image Analysis, IbPRIA 2019, pp. 147–158, Springer, Madrid, 2019, ISBN: 978-3-030-31321-0.
Links | BibTeX | Tags: HispaMus
@inproceedings{k419,
title = {Approaching End-to-end Optical Music Recognition for Homophonic Scores},
author = {M. Alfaro-Contreras and J. Calvo-Zaragoza and J. M. Iñesta},
url = {https://link.springer.com/chapter/10.1007/978-3-030-31321-0_13},
isbn = {978-3-030-31321-0},
year = {2019},
date = {2019-08-01},
urldate = {2019-08-01},
booktitle = {Iberian Conference on Pattern Recognition and Image Analysis, IbPRIA 2019},
pages = {147--158},
publisher = {Springer},
address = {Madrid},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Mateiu, T. N.; Gallego, A. J.; Calvo-Zaragoza, J.
Domain Adaptation for Handwritten Symbol Recognition: A Case of Study in Old Music Manuscripts Proceedings Article
In: Iberian Conference on Pattern Recognition and Image Analysis, IbPRIA, pp. 135–146, Springer International Publishing, Madrid, Spain, 2019, ISBN: 978-3-030-31321-0.
Abstract | BibTeX | Tags: HispaMus
@inproceedings{k410,
title = {Domain Adaptation for Handwritten Symbol Recognition: A Case of Study in Old Music Manuscripts},
author = {T. N. Mateiu and A. J. Gallego and J. Calvo-Zaragoza},
isbn = {978-3-030-31321-0},
year = {2019},
date = {2019-07-01},
urldate = {2019-07-01},
booktitle = {Iberian Conference on Pattern Recognition and Image Analysis, IbPRIA},
pages = {135--146},
publisher = {Springer International Publishing},
address = {Madrid, Spain},
abstract = {The existence of a large amount of untranscripted music manuscripts has caused initiatives that use Machine Learning (ML) for Optical Music Recognition, in order to efficiently transcribe the music sources into a machine-readable format. Although most music manuscript are similar in nature, they inevitably vary from one another. This fact can negatively influence the complexity of the classification task because most ML models fail to transfer their knowledge from one domain to another, thereby requiring learning from scratch on new domains after manually labeling new data. This work studies the ability of a Domain Adversarial Neural Network for domain adaptation in the context of classifying handwritten music symbols. The main idea is to exploit the knowledge of a specific manuscript to classify symbols from different (unlabeled) manuscripts. The reported results are promising, obtaining a substantial improvement over a conventional Convolutional Neural Network approach, which can be used as a basis for future research.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
The existence of a large amount of untranscripted music manuscripts has caused initiatives that use Machine Learning (ML) for Optical Music Recognition, in order to efficiently transcribe the music sources into a machine-readable format. Although most music manuscript are similar in nature, they inevitably vary from one another. This fact can negatively influence the complexity of the classification task because most ML models fail to transfer their knowledge from one domain to another, thereby requiring learning from scratch on new domains after manually labeling new data. This work studies the ability of a Domain Adversarial Neural Network for domain adaptation in the context of classifying handwritten music symbols. The main idea is to exploit the knowledge of a specific manuscript to classify symbols from different (unlabeled) manuscripts. The reported results are promising, obtaining a substantial improvement over a conventional Convolutional Neural Network approach, which can be used as a basis for future research. Gallego, A. J.; Pertusa, A.; Bernabeu, M.
Multi-Label Logo Classification using Convolutional Neural Networks Proceedings Article
In: Iberian Conference on Pattern Recognition and Image Analysis, IbPRIA, pp. 485–497, Springer International Publishing, Madrid, Spain, 2019, ISBN: 978-3-030-31332-6.
Abstract | BibTeX | Tags: HispaMus
@inproceedings{k411,
title = {Multi-Label Logo Classification using Convolutional Neural Networks},
author = {A. J. Gallego and A. Pertusa and M. Bernabeu},
isbn = {978-3-030-31332-6},
year = {2019},
date = {2019-07-01},
urldate = {2019-07-01},
booktitle = {Iberian Conference on Pattern Recognition and Image Analysis, IbPRIA},
pages = {485--497},
publisher = {Springer International Publishing},
address = {Madrid, Spain},
abstract = {The classification of logos is a particular case within computer vision since they have their own characteristics. Logos can contain only text, iconic images or a combination of both, and they usually include figurative symbols designed by experts that vary substantially besides they may share the same semantics. This work presents a method for multi-label classification and retrieval of logo images. For this, Convolutional Neural Networks (CNN) are trained to classify logos from the European Union TradeMark (EUTM) dataset according to their colors, shapes, sectors and figurative designs. An auto-encoder is also trained to learn representations of the input images. Once trained, the neural codes from the last convolutional layers in the CNN and the central layer of the auto-encoder can be used to perform similarity search through kNN, allowing us to obtain the most similar logos based on their color, shape, sector, figurative elements, overall features, or a weighted combination of them provided by the user. To the best of our knowledge, this is the first multi-label classification method for logos, and the only one that allows retrieving a ranking of images with these criteria provided by the user.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
The classification of logos is a particular case within computer vision since they have their own characteristics. Logos can contain only text, iconic images or a combination of both, and they usually include figurative symbols designed by experts that vary substantially besides they may share the same semantics. This work presents a method for multi-label classification and retrieval of logo images. For this, Convolutional Neural Networks (CNN) are trained to classify logos from the European Union TradeMark (EUTM) dataset according to their colors, shapes, sectors and figurative designs. An auto-encoder is also trained to learn representations of the input images. Once trained, the neural codes from the last convolutional layers in the CNN and the central layer of the auto-encoder can be used to perform similarity search through kNN, allowing us to obtain the most similar logos based on their color, shape, sector, figurative elements, overall features, or a weighted combination of them provided by the user. To the best of our knowledge, this is the first multi-label classification method for logos, and the only one that allows retrieving a ranking of images with these criteria provided by the user. Nuñez-Alcover, Alicia; León, P. J. Ponce; Calvo-Zaragoza, J.
Glyph and Position Classification of Music Symbols in Early Music Manuscripts Proceedings Article
In: Morales, A.; Fierrez, J.; Sánchez, J. Salvador; Ribeiro, B. (Ed.): Proc. of the 9th Iberian Conference on Pattern Recognition and Image Analysis, LNCS vol. 11867, pp. 159-168, AERFAI, APRP Springer, Madrid, Spain, 2019, ISBN: 978-3-030-31331-9.
Abstract | Links | BibTeX | Tags: HispaMus
@inproceedings{k420,
title = {Glyph and Position Classification of Music Symbols in Early Music Manuscripts},
author = {Alicia Nuñez-Alcover and P. J. Ponce León and J. Calvo-Zaragoza},
editor = {A. Morales and J. Fierrez and J. Salvador Sánchez and B. Ribeiro},
url = {https://grfia.dlsi.ua.es/repositori/grfia/pubs/420/IbPRIA2019_paper_91-1.pdf},
isbn = {978-3-030-31331-9},
year = {2019},
date = {2019-07-01},
booktitle = {Proc. of the 9th Iberian Conference on Pattern Recognition and Image Analysis, LNCS vol. 11867},
pages = {159-168},
publisher = {Springer},
address = {Madrid, Spain},
organization = {AERFAI, APRP},
abstract = {Optical Music Recognition is a field of research that automates the reading of musical scores so as to transcribe their content into a structured digital format. When dealing with music manuscripts, the traditional workflow establishes separate stages of detection and classification of musical symbols. In the latter, most of the research has focused on detecting musical glyphs, ignoring that the meaning of a musical symbol is defined by two components: its glyph and its position within the staff. In this paper we study how to perform both glyph and position classification of handwritten musical symbols in early music manuscripts written in white Mensural notation, a common notation system used for the most part of the XVI and XVII centuries. We make use of Convolutional Neural Networks as the classification method, and we tested several alternatives such as using independent models for each component, combining label spaces, or using both multi-input and multi-output models. Our results on early music manuscripts provide insights about the effectiveness and efficiency of each approach.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Optical Music Recognition is a field of research that automates the reading of musical scores so as to transcribe their content into a structured digital format. When dealing with music manuscripts, the traditional workflow establishes separate stages of detection and classification of musical symbols. In the latter, most of the research has focused on detecting musical glyphs, ignoring that the meaning of a musical symbol is defined by two components: its glyph and its position within the staff. In this paper we study how to perform both glyph and position classification of handwritten musical symbols in early music manuscripts written in white Mensural notation, a common notation system used for the most part of the XVI and XVII centuries. We make use of Convolutional Neural Networks as the classification method, and we tested several alternatives such as using independent models for each component, combining label spaces, or using both multi-input and multi-output models. Our results on early music manuscripts provide insights about the effectiveness and efficiency of each approach. Bustos, A.
Extraction of medical knowledge from clinical reports and chest x-rays using machine learning techniques PhD Thesis
2019.
BibTeX | Tags:
@phdthesis{k459,
title = {Extraction of medical knowledge from clinical reports and chest x-rays using machine learning techniques},
author = {A. Bustos},
editor = {A. Pertusa},
year = {2019},
date = {2019-07-01},
urldate = {2019-07-01},
organization = {University of Alicante},
keywords = {},
pubstate = {published},
tppubtype = {phdthesis}
}
Gallego, A. J.; Gil, P.; Pertusa, A.; Fisher, R. B.
Semantic Segmentation of SLAR Imagery with Convolutional LSTM Selectional AutoEncoders Journal Article
In: Remote Sensing, vol. 11, no. 12, pp. 1-22, 2019.
@article{k412,
title = {Semantic Segmentation of SLAR Imagery with Convolutional LSTM Selectional AutoEncoders},
author = {A. J. Gallego and P. Gil and A. Pertusa and R. B. Fisher},
year = {2019},
date = {2019-06-01},
journal = {Remote Sensing},
volume = {11},
number = {12},
pages = {1-22},
abstract = {We present a method to detect maritime oil spills from Side-Looking Airborne Radar (SLAR) sensors mounted on aircraft in order to enable a quick response of emergency services when an oil spill occurs. The proposed approach introduces a new type of neural architecture named Convolutional Long Short Term Memory Selectional AutoEncoders (CMSAE) which allows the simultaneous segmentation of multiple classes such as coast, oil spill and ships. Unlike previous works using full SLAR images, in this work only a few scanlines from the beam-scanning of radar are needed to perform the detection. The main objective is to develop a method that performs accurate segmentation using only the current and previous sensor information, in order to return a real-time response during the flight. The proposed architecture uses a series of CMSAE networks to process in parallel each of the objectives defined as different classes. The output of these networks are given to a machine learning classifier to perform the final detection. Results show that the proposed approach can reliably detect oil spills and other maritime objects in SLAR sequences, outperforming the accuracy of previous state-of-the-art methods and with a response time of only 0.76 s.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
We present a method to detect maritime oil spills from Side-Looking Airborne Radar (SLAR) sensors mounted on aircraft in order to enable a quick response of emergency services when an oil spill occurs. The proposed approach introduces a new type of neural architecture named Convolutional Long Short Term Memory Selectional AutoEncoders (CMSAE) which allows the simultaneous segmentation of multiple classes such as coast, oil spill and ships. Unlike previous works using full SLAR images, in this work only a few scanlines from the beam-scanning of radar are needed to perform the detection. The main objective is to develop a method that performs accurate segmentation using only the current and previous sensor information, in order to return a real-time response during the flight. The proposed architecture uses a series of CMSAE networks to process in parallel each of the objectives defined as different classes. The output of these networks are given to a machine learning classifier to perform the final detection. Results show that the proposed approach can reliably detect oil spills and other maritime objects in SLAR sequences, outperforming the accuracy of previous state-of-the-art methods and with a response time of only 0.76 s. Calvo-Zaragoza, J.; Rico-Juan, J. R.; Gallego, A. J.
Ensemble classification from deep predictions with test data augmentation Journal Article
In: Soft Computing, 2019, ISSN: 1433-7479.
@article{k409,
title = {Ensemble classification from deep predictions with test data augmentation},
author = {J. Calvo-Zaragoza and J. R. Rico-Juan and A. J. Gallego},
issn = {1433-7479},
year = {2019},
date = {2019-04-01},
urldate = {2019-04-01},
journal = {Soft Computing},
abstract = {Data augmentation has become a standard step to improve the predictive power and robustness of convolutional neural networks by means of the synthetic generation of new samples depicting different deformations. This step has been traditionally considered to improve the network at the training stage. In this work, however, we study the use of data augmentation at classification time. That is, the test sample is augmented, following the same procedure considered for training, and the decision is taken with an ensemble prediction over all these samples. We present comprehensive experimentation with several datasets and ensemble decisions, considering a rather generic data augmentation procedure. Our results show that performing this step is able to boost the original classification, even when the room for improvement is limited.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Data augmentation has become a standard step to improve the predictive power and robustness of convolutional neural networks by means of the synthetic generation of new samples depicting different deformations. This step has been traditionally considered to improve the network at the training stage. In this work, however, we study the use of data augmentation at classification time. That is, the test sample is augmented, following the same procedure considered for training, and the decision is taken with an ensemble prediction over all these samples. We present comprehensive experimentation with several datasets and ensemble decisions, considering a rather generic data augmentation procedure. Our results show that performing this step is able to boost the original classification, even when the room for improvement is limited. Rizo, D.; Marsden, A.
An MEI-based standard encoding for hierarchical music analyses Journal Article
In: International Journal on Digital Libraries, vol. 20, no. 1, pp. 93–105, 2019, ISSN: 1432-5012.
@article{k402,
title = {An MEI-based standard encoding for hierarchical music analyses},
author = {D. Rizo and A. Marsden},
issn = {1432-5012},
year = {2019},
date = {2019-03-01},
journal = {International Journal on Digital Libraries},
volume = {20},
number = {1},
pages = {93--105},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {article}
}
Calvo-Zaragoza, J.; Gallego, A. J.
A selectional auto-encoder approach for document image binarization Journal Article
In: Pattern Recognition, vol. 86, pp. 37-47, 2019, ISSN: 0031-3203.
Abstract | Links | BibTeX | Tags: GRE16-14
@article{k395,
title = {A selectional auto-encoder approach for document image binarization},
author = {J. Calvo-Zaragoza and A. J. Gallego},
url = {https://grfia.dlsi.ua.es/repositori/grfia/pubs/395/1706.10241.pdf},
issn = {0031-3203},
year = {2019},
date = {2019-01-01},
journal = {Pattern Recognition},
volume = {86},
pages = {37-47},
abstract = {Binarization plays a key role in the automatic information retrieval from document images. This process is usually performed in the first stages of document analysis systems, and serves as a basis for subsequent steps. Hence it has to be robust in order to allow the full analysis workflow to be successful. Several methods for document image binarization have been proposed so far, most of which are based on hand-crafted image processing strategies. Recently, Convolutional Neural Networks have shown an amazing performance in many disparate duties related to computer vision. In this paper we discuss the use of convolutional auto-encoders devoted to learning an end-to-end map from an input image to its selectional output, in which activations indicate the likelihood of pixels to be either foreground or background. Once trained, documents can therefore be binarized by parsing them through the model and applying a global threshold. This approach has proven to outperform existing binarization strategies in a number of document types.},
keywords = {GRE16-14},
pubstate = {published},
tppubtype = {article}
}
Binarization plays a key role in the automatic information retrieval from document images. This process is usually performed in the first stages of document analysis systems, and serves as a basis for subsequent steps. Hence it has to be robust in order to allow the full analysis workflow to be successful. Several methods for document image binarization have been proposed so far, most of which are based on hand-crafted image processing strategies. Recently, Convolutional Neural Networks have shown an amazing performance in many disparate duties related to computer vision. In this paper we discuss the use of convolutional auto-encoders devoted to learning an end-to-end map from an input image to its selectional output, in which activations indicate the likelihood of pixels to be either foreground or background. Once trained, documents can therefore be binarized by parsing them through the model and applying a global threshold. This approach has proven to outperform existing binarization strategies in a number of document types. Gallego, A. J.; Pertusa, A.; Gil, P.; Fisher, R. B.
Detection of bodies in maritime rescue operations using unmanned aerial vehicles with multispectral cameras Journal Article
In: Journal of Field Robotics, vol. 36, no. 4, pp. 782-796, 2019.
BibTeX | Tags:
@article{k401,
title = {Detection of bodies in maritime rescue operations using unmanned aerial vehicles with multispectral cameras},
author = {A. J. Gallego and A. Pertusa and P. Gil and R. B. Fisher},
year = {2019},
date = {2019-01-01},
journal = {Journal of Field Robotics},
volume = {36},
number = {4},
pages = {782-796},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Alashhab, S.; Gallego, A. J.; Lozano, M. Á.
Hand Gesture Detection with Convolutional Neural Networks Proceedings Article
In: Distributed Computing and Artificial Intelligence, 15th International Conference, pp. 45–52, Springer International Publishing, 2019, ISBN: 978-3-319-94649-8.
@inproceedings{k405,
title = {Hand Gesture Detection with Convolutional Neural Networks},
author = {S. Alashhab and A. J. Gallego and M. Á. Lozano},
isbn = {978-3-319-94649-8},
year = {2019},
date = {2019-01-01},
urldate = {2019-01-01},
booktitle = {Distributed Computing and Artificial Intelligence, 15th International Conference},
pages = {45--52},
publisher = {Springer International Publishing},
abstract = {In this paper, we present a method for locating and recognizing hand gestures from images, based on Deep Learning. Our goal is to provide an intuitive and accessible way to interact with Computer Vision-based mobile applications aimed to assist visually impaired people (e.g. pointing a finger at an object in a real scene to zoom in for a close-up of the pointed object). Initially, we have defined different hand gestures that can be assigned to different actions. After that, we have created a database containing images corresponding to these gestures. Lastly, this database has been used to train Neural Networks with different topologies (testing different input sizes, weight initialization, and data augmentation process). In our experiments, we have obtained high accuracies both in localization (96%-100%) and in recognition (99.45%) with Networks that are appropriate to be ported to mobile devices.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
In this paper, we present a method for locating and recognizing hand gestures from images, based on Deep Learning. Our goal is to provide an intuitive and accessible way to interact with Computer Vision-based mobile applications aimed to assist visually impaired people (e.g. pointing a finger at an object in a real scene to zoom in for a close-up of the pointed object). Initially, we have defined different hand gestures that can be assigned to different actions. After that, we have created a database containing images corresponding to these gestures. Lastly, this database has been used to train Neural Networks with different topologies (testing different input sizes, weight initialization, and data augmentation process). In our experiments, we have obtained high accuracies both in localization (96%-100%) and in recognition (99.45%) with Networks that are appropriate to be ported to mobile devices. Garay-Maestre, U.; Gallego, A. J.; Calvo-Zaragoza, J.
Data Augmentation via Variational Auto-Encoders Proceedings Article
In: Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications, pp. 29–37, 2019, ISBN: 978-3-030-13469-3.
Abstract | BibTeX | Tags: HispaMus
@inproceedings{k406,
title = {Data Augmentation via Variational Auto-Encoders},
author = {U. Garay-Maestre and A. J. Gallego and J. Calvo-Zaragoza},
isbn = {978-3-030-13469-3},
year = {2019},
date = {2019-01-01},
urldate = {2019-01-01},
booktitle = {Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications},
pages = {29--37},
abstract = {Data augmentation is a widely considered technique to improve the performance of Convolutional Neural Networks during training. This step consists in synthetically generate new labeled data by perturbing the samples of the training set, which is expected to provide more robustness to the learning process. The problem is that the augmentation procedure has to be adjusted manually because the perturbations considered must make sense for the task at issue. In this paper we propose the use of Variational Auto-Encoders (VAEs) to generate new synthetic samples, instead of resorting to heuristic strategies. VAEs are powerful generative models that learn a parametric latent space of the input domain from which new samples can be generated. In our experiments over the well-known MNIST dataset, the data augmentation by VAEs improves the base results, yet to a lesser extent of that obtained by a well-adjusted conventional data augmentation. However, the combination of both conventional and VAE-guided data augmentations outperforms all the results, thereby demonstrating the goodness of our proposal.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Data augmentation is a widely considered technique to improve the performance of Convolutional Neural Networks during training. This step consists in synthetically generate new labeled data by perturbing the samples of the training set, which is expected to provide more robustness to the learning process. The problem is that the augmentation procedure has to be adjusted manually because the perturbations considered must make sense for the task at issue. In this paper we propose the use of Variational Auto-Encoders (VAEs) to generate new synthetic samples, instead of resorting to heuristic strategies. VAEs are powerful generative models that learn a parametric latent space of the input domain from which new samples can be generated. In our experiments over the well-known MNIST dataset, the data augmentation by VAEs improves the base results, yet to a lesser extent of that obtained by a well-adjusted conventional data augmentation. However, the combination of both conventional and VAE-guided data augmentations outperforms all the results, thereby demonstrating the goodness of our proposal.
2021
Calvo-Zaragoza, J.; Rizo, D.; Iñesta, J. M.
Reconocimiento Óptico de Partituras (OMR) aplicado al Fonde de Música Tradicional IMF-CSIC Book Chapter
In: Gambero-Ustárroz, M.; Ros-Fábregas, E. (Ed.): Musicología en Web. Patrimonio musical y Humanidades Digitales, Chapter 4, pp. 87-109, Edition Reichenberger, 2021, ISBN: 978-3-967280-14-2.
@inbook{k470,
title = {Reconocimiento Óptico de Partituras (OMR) aplicado al Fonde de Música Tradicional IMF-CSIC},
author = {J. Calvo-Zaragoza and D. Rizo and J. M. Iñesta},
editor = {M. Gambero-Ustárroz and E. Ros-Fábregas},
isbn = {978-3-967280-14-2},
year = {2021},
date = {2021-01-01},
booktitle = {Musicología en Web. Patrimonio musical y Humanidades Digitales},
pages = {87-109},
publisher = {Edition Reichenberger},
chapter = {4},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inbook}
}
Ortega-Bastida, J.; Gallego, A. J.; Rico-Juan, J. R.; Albarrán, P.
A multimodal approach for regional GDP prediction using social media activity and historical information Journal Article
In: Applied Soft Computing, pp. 107693, 2021, ISSN: 1568-4946.
@article{k471,
title = {A multimodal approach for regional GDP prediction using social media activity and historical information},
author = {J. Ortega-Bastida and A. J. Gallego and J. R. Rico-Juan and P. Albarrán},
issn = {1568-4946},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
journal = {Applied Soft Computing},
pages = {107693},
abstract = {This work proposes a multimodal approach with which to predict the regional Gross Domestic Product (GDP) by combining historical GDP values with the embodied information in Twitter messages concerning the current economic condition. This proposal is of great interest, since it delivers forecasts at higher frequencies than both the official statistics (published only annually at the regional level in Spain) and the existing unofficial quarterly predictions (which rely on economic indicators that are available only after months of delay). The proposed method is based on a two-stage architecture. In the first stage, a multi-task autoencoder is initially used to obtain a GDP-related representation of tweets, which are then filtered to remove outliers and to obtain the GDP prediction from the consensus of opinions. In a second stage, this result is combined with the historical GDP values of the region using a multimodal network. The method is evaluated in four different regions of Spain using the tweets written by the most relevant economists, politicians, newspapers and institutions in each one. The results show that our approach successfully learns the evolution of the GDP using only historical information and tweets, thus making it possible to provide earlier forecasts about the regional GDP. This method also makes it possible to establish which the most or least influential opinions regarding this prediction are. As an additional exercise, we have assessed how well our method predicted the effect of the COVID-19 pandemic.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
This work proposes a multimodal approach with which to predict the regional Gross Domestic Product (GDP) by combining historical GDP values with the embodied information in Twitter messages concerning the current economic condition. This proposal is of great interest, since it delivers forecasts at higher frequencies than both the official statistics (published only annually at the regional level in Spain) and the existing unofficial quarterly predictions (which rely on economic indicators that are available only after months of delay). The proposed method is based on a two-stage architecture. In the first stage, a multi-task autoencoder is initially used to obtain a GDP-related representation of tweets, which are then filtered to remove outliers and to obtain the GDP prediction from the consensus of opinions. In a second stage, this result is combined with the historical GDP values of the region using a multimodal network. The method is evaluated in four different regions of Spain using the tweets written by the most relevant economists, politicians, newspapers and institutions in each one. The results show that our approach successfully learns the evolution of the GDP using only historical information and tweets, thus making it possible to provide earlier forecasts about the regional GDP. This method also makes it possible to establish which the most or least influential opinions regarding this prediction are. As an additional exercise, we have assessed how well our method predicted the effect of the COVID-19 pandemic.
López-Gutiérrez, J. C.; Valero-Mas, J. J.; Castellanos, F. J.; Calvo-Zaragoza, J.
Data Augmentation for End-to-End Optical Music Recognition Proceedings Article
In: Proceedings of the 14th IAPR International Workshop on Graphics Recognition (GREC), pp. 59-73, Springer, 2021.
BibTeX | Tags: GV/2020/030
@inproceedings{k473,
title = {Data Augmentation for End-to-End Optical Music Recognition},
author = {J. C. López-Gutiérrez and J. J. Valero-Mas and F. J. Castellanos and J. Calvo-Zaragoza},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
booktitle = {Proceedings of the 14th IAPR International Workshop on Graphics Recognition (GREC)},
pages = {59-73},
publisher = {Springer},
keywords = {GV/2020/030},
pubstate = {published},
tppubtype = {inproceedings}
}
Alfaro-Contreras, M.; Valero-Mas, J. J.; Iñesta, J. M.
Neural architectures for exploiting the components of Agnostic Notation in Optical Music Recognition Proceedings Article
In: Proc. of the 3rd International Workshop on Reading Music Systems, pp. 13-17, 2021.
BibTeX | Tags:
@inproceedings{k474,
title = {Neural architectures for exploiting the components of Agnostic Notation in Optical Music Recognition},
author = {M. Alfaro-Contreras and J. J. Valero-Mas and J. M. Iñesta},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
booktitle = {Proc. of the 3rd International Workshop on Reading Music Systems},
pages = {13-17},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Castellanos, F. J.; Gallego, A. J.; Calvo-Zaragoza, J.
Unsupervised Domain Adaptation for Document Analysis of Music Score Images Proceedings Article
In: Proc. of the 22nd International Society for Music Information Retrieval Conference, 2021.
@inproceedings{k475,
title = {Unsupervised Domain Adaptation for Document Analysis of Music Score Images},
author = {F. J. Castellanos and A. J. Gallego and J. Calvo-Zaragoza},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
booktitle = {Proc. of the 22nd International Society for Music Information Retrieval Conference},
keywords = {GRE19-04},
pubstate = {published},
tppubtype = {inproceedings}
}
Castellanos, F. J.; Valero-Mas, J. J.; Calvo-Zaragoza, J.
Prototype Generation in the String Space via Approximate Median for Data Reduction in Nearest Neighbor classification Journal Article
In: Soft Computing, vol. 25, 2021, ISSN: 15403-15415.
BibTeX | Tags: GRE19-04, HispaMus
@article{k476,
title = {Prototype Generation in the String Space via Approximate Median for Data Reduction in Nearest Neighbor classification},
author = {F. J. Castellanos and J. J. Valero-Mas and J. Calvo-Zaragoza},
issn = {15403-15415},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
journal = {Soft Computing},
volume = {25},
keywords = {GRE19-04, HispaMus},
pubstate = {published},
tppubtype = {article}
}
Ríos-Vila, A.; Rizo, D.; Calvo-Zaragoza, J.
Complete Optical Music Recognition via Agnostic Transcription and Machine Translation Proceedings Article
In: 16th International Conference on Document Analysis and Recognition, pp. 661-675, 2021.
BibTeX | Tags: GV/2020/030
@inproceedings{k477,
title = {Complete Optical Music Recognition via Agnostic Transcription and Machine Translation},
author = {A. Ríos-Vila and D. Rizo and J. Calvo-Zaragoza},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
booktitle = {16th International Conference on Document Analysis and Recognition},
pages = {661-675},
keywords = {GV/2020/030},
pubstate = {published},
tppubtype = {inproceedings}
}
Mas-Candela, E.; Alfaro-Contreras, M.; Calvo-Zaragoza, J.
Sequential Next-Symbol Prediction for Optical Music Recognition Proceedings Article
In: 16th International Conference on Document Analysis and Recognition, pp. 708-722, 2021, ISBN: 978-3-030-86334-0.
Links | BibTeX | Tags: GV/2020/030
@inproceedings{k478,
title = {Sequential Next-Symbol Prediction for Optical Music Recognition},
author = {E. Mas-Candela and M. Alfaro-Contreras and J. Calvo-Zaragoza},
url = {https://link.springer.com/chapter/10.1007/978-3-030-86334-0_46},
isbn = {978-3-030-86334-0},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
booktitle = {16th International Conference on Document Analysis and Recognition},
pages = {708-722},
keywords = {GV/2020/030},
pubstate = {published},
tppubtype = {inproceedings}
}
Garrido-Munoz, C.; Sánchez-Hernández, A.; Castellanos, F. J.; Calvo-Zaragoza, J.
Domain Adaptation for Document Image Binarization via Domain Classification Proceedings Article
In: Tallón-Ballesteros, A. J. (Ed.): Frontiers in Artificial Intelligence and Applications, pp. 569-582, IOS Press, 2021, ISBN: 978-1-64368-224-2.
BibTeX | Tags: GRE19-04, GV/2020/030
@inproceedings{k480,
title = {Domain Adaptation for Document Image Binarization via Domain Classification},
author = {C. Garrido-Munoz and A. Sánchez-Hernández and F. J. Castellanos and J. Calvo-Zaragoza},
editor = {A. J. Tallón-Ballesteros},
isbn = {978-1-64368-224-2},
year = {2021},
date = {2021-01-01},
booktitle = {Frontiers in Artificial Intelligence and Applications},
pages = {569-582},
publisher = {IOS Press},
chapter = {-},
keywords = {GRE19-04, GV/2020/030},
pubstate = {published},
tppubtype = {inproceedings}
}
Ortega-Bastida, J.
Aproximación de soluciones multimodales para aplicaciones Deep Learning PhD Thesis
2021.
BibTeX | Tags:
@phdthesis{k507,
title = {Aproximación de soluciones multimodales para aplicaciones Deep Learning},
author = {J. Ortega-Bastida},
editor = {J. Ramón Rico-Juan and A. J. Gallego},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
organization = {Universidad de Alicante},
keywords = {},
pubstate = {published},
tppubtype = {phdthesis}
}
Cuevas-Velasquez, H.; Gallego, A. J.; Fisher, R. B.
Two Heads are Better than One: Geometric-Latent Attention for Point Cloud Classification and Segmentation Proceedings Article
In: The 32nd British Machine Vision Conference (BMVC), 2021.
BibTeX | Tags:
@inproceedings{k513,
title = {Two Heads are Better than One: Geometric-Latent Attention for Point Cloud Classification and Segmentation},
author = {H. Cuevas-Velasquez and A. J. Gallego and R. B. Fisher},
year = {2021},
date = {2021-01-01},
booktitle = {The 32nd British Machine Vision Conference (BMVC)},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
2020
Bustos, A.; Pertusa, A.; Salinas, J. M.; Iglesia-Vayá, M.
PadChest: A large chest x-ray image dataset with multi-label annotated reports Journal Article
In: Medical Image Analysis, vol. 66, pp. 101797, 2020, ISSN: 1361-8415.
@article{k404,
title = {PadChest: A large chest x-ray image dataset with multi-label annotated reports},
author = {A. Bustos and A. Pertusa and J. M. Salinas and M. Iglesia-Vayá},
issn = {1361-8415},
year = {2020},
date = {2020-12-01},
journal = {Medical Image Analysis},
volume = {66},
pages = {101797},
abstract = {We present a labeled large-scale, high resolution chest x-ray dataset for the automated exploration of medical images along with their associated reports. This dataset includes more than 160,000 images obtained from 67,000 patients that were interpreted and reported by radiologists at San Juan Hospital (Spain) from 2009 to 2017, covering six different position views and additional information on image acquisition and patient demography. The reports were labeled with 174 different radiographic findings, 19 differential diagnoses and 104 anatomic locations organized as a hierarchical taxonomy and mapped onto standard Unified Medical Language System (UMLS) terminology. Of these reports, 27% were manually annotated by trained physicians and the remaining set was labeled using a supervised method based on a recurrent neural network with attention mechanisms. The labels generated were then validated in an independent test set achieving a 0.93 Micro-F1 score. To the best of our knowledge, this is one of the largest public chest x-ray databases suitable for training supervised models concerning radiographs, and the first to contain radiographic reports in Spanish. The PadChest dataset can be downloaded from http://bimcv.cipf.es/bimcv-projects/padchest/.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
We present a labeled large-scale, high resolution chest x-ray dataset for the automated exploration of medical images along with their associated reports. This dataset includes more than 160,000 images obtained from 67,000 patients that were interpreted and reported by radiologists at San Juan Hospital (Spain) from 2009 to 2017, covering six different position views and additional information on image acquisition and patient demography. The reports were labeled with 174 different radiographic findings, 19 differential diagnoses and 104 anatomic locations organized as a hierarchical taxonomy and mapped onto standard Unified Medical Language System (UMLS) terminology. Of these reports, 27% were manually annotated by trained physicians and the remaining set was labeled using a supervised method based on a recurrent neural network with attention mechanisms. The labels generated were then validated in an independent test set achieving a 0.93 Micro-F1 score. To the best of our knowledge, this is one of the largest public chest x-ray databases suitable for training supervised models concerning radiographs, and the first to contain radiographic reports in Spanish. The PadChest dataset can be downloaded from http://bimcv.cipf.es/bimcv-projects/padchest/.
Rico-Juan, J. R.; Valero-Mas, J. J.; Iñesta, J. M.
Bounding Edit Distance for similarity-based sequence classification on Structural Pattern Recognition Journal Article
In: Applied Soft Computing, vol. 97 (Part A), 2020.
Abstract | BibTeX | Tags: HispaMus
@article{k451,
title = {Bounding Edit Distance for similarity-based sequence classification on Structural Pattern Recognition},
author = {J. R. Rico-Juan and J. J. Valero-Mas and J. M. Iñesta},
year = {2020},
date = {2020-12-01},
journal = {Applied Soft Computing},
volume = {97 (Part A)},
abstract = {Pattern Recognition tasks in the structural domain generally exhibit high accuracy results, but their time efficiency is quite low. Furthermore, this low performance is more pronounced when dealing with instance-based classifiers, since, for each query, the entire corpus must be evaluated to find the closest prototype. In this work we address this efficiency issue for the Nearest Neighbor classifier when data are encoded as two-dimensional code sequences, and more precisely strings and sequences of vectors. For this, a set of bounds is proposed in the distance metric that avoid the calculation of unnecessary distances. Results obtained prove the effectiveness of the proposal as it reduces the classification time in percentages between 80% and 90% for string representations and between 60% and 80% for data codified as sequences of vectors with respect to their corresponding non-optimized version of the classifier.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {article}
}
Pattern Recognition tasks in the structural domain generally exhibit high accuracy results, but their time efficiency is quite low. Furthermore, this low performance is more pronounced when dealing with instance-based classifiers, since, for each query, the entire corpus must be evaluated to find the closest prototype. In this work we address this efficiency issue for the Nearest Neighbor classifier when data are encoded as two-dimensional code sequences, and more precisely strings and sequences of vectors. For this, a set of bounds is proposed in the distance metric that avoid the calculation of unnecessary distances. Results obtained prove the effectiveness of the proposal as it reduces the classification time in percentages between 80% and 90% for string representations and between 60% and 80% for data codified as sequences of vectors with respect to their corresponding non-optimized version of the classifier.
Saborit-Torres, J. M.; Saenz-Gamboa, J. J.; Montell, J. À.; Salinas, J. M.; Gómez, J. A.; Stefan, I.; Caparrós, M.; García-García, F.; Domenech, J.; Manjón, J. V.; Rojas, G.; Pertusa, A.; Bustos, A.; González, G.; Galant, J.; Iglesia-Vayá, M.
Medical imaging data structure extended to multiple modalities and anatomical regions Journal Article
In: ArXiv:2010.00434, 2020.
@article{k452,
title = {Medical imaging data structure extended to multiple modalities and anatomical regions},
author = {J. M. Saborit-Torres and J. J. Saenz-Gamboa and J. À. Montell and J. M. Salinas and J. A. Gómez and I. Stefan and M. Caparrós and F. García-García and J. Domenech and J. V. Manjón and G. Rojas and A. Pertusa and A. Bustos and G. González and J. Galant and M. Iglesia-Vayá},
year = {2020},
date = {2020-10-01},
journal = {ArXiv:2010.00434},
abstract = {Brain Imaging Data Structure (BIDS) allows the user to organise brain imaging data into a clear and easy standard directory structure. BIDS is widely supported by the scientific community and is considered a powerful standard for management. The original BIDS is limited to images or data related to the brain. Medical Imaging Data Structure (MIDS) was therefore conceived with the objective of extending this methodology to other anatomical regions and other types of imaging systems in these areas.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Brain Imaging Data Structure (BIDS) allows the user to organise brain imaging data into a clear and easy standard directory structure. BIDS is widely supported by the scientific community and is considered a powerful standard for management. The original BIDS is limited to images or data related to the brain. Medical Imaging Data Structure (MIDS) was therefore conceived with the objective of extending this methodology to other anatomical regions and other types of imaging systems in these areas.
Ríos-Vila, A.; Calvo-Zaragoza, J.; Rizo, D.
Evaluating Simultaneous Recognition and Encoding for Optical Music Recognition Proceedings Article
In: DLfM 2020: 7th International Conference on Digital Libraries for Musicology, pp. 10-17, Association for Computing Machinery, 2020, ISBN: 978-1-4503-8760-6.
@inproceedings{k456,
title = {Evaluating Simultaneous Recognition and Encoding for Optical Music Recognition},
author = {A. Ríos-Vila and J. Calvo-Zaragoza and D. Rizo},
isbn = {978-1-4503-8760-6},
year = {2020},
date = {2020-10-01},
booktitle = {DLfM 2020: 7th International Conference on Digital Libraries for Musicology},
journal = {DLfM 2020: 7th International Conference on Digital Libraries for Musicology},
pages = {10-17},
publisher = {Association for Computing Machinery},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Castellanos, F. J.; Calvo-Zaragoza, J.; Iñesta, J. M.
A neural approach for full-page Optical Music Recognition of mensural documents Proceedings Article
In: Proceedings of the 21st International Society for Music Information Retrieval Conference, pp. 558-565, Montréal, Canada, 2020, ISBN: 978-0-9813537-0-8.
@inproceedings{k457,
title = {A neural approach for full-page Optical Music Recognition of mensural documents},
author = {F. J. Castellanos and J. Calvo-Zaragoza and J. M. Iñesta},
isbn = {978-0-9813537-0-8},
year = {2020},
date = {2020-10-01},
booktitle = {Proceedings of the 21st International Society for Music Information Retrieval Conference},
pages = {558-565},
address = {Montréal, Canada},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Castellanos, F. J.; Calvo-Zaragoza, J.; Iñesta, J. M.
Two-step neural cross-domain experiments forfull-page recognition of Mensural documents Proceedings Article
In: Proc. of the 13th International Workshop on Machine Learning and Music, pp. 56–60, MML 2020 13th International Workshop on Machine Learning and Music, 2020.
BibTeX | Tags: GRE19-04, HispaMus
@inproceedings{k448,
title = {Two-step neural cross-domain experiments forfull-page recognition of Mensural documents},
author = {F. J. Castellanos and J. Calvo-Zaragoza and J. M. Iñesta},
year = {2020},
date = {2020-09-01},
booktitle = {Proc. of the 13th International Workshop on Machine Learning and Music},
pages = {56--60},
publisher = {13th International Workshop on Machine Learning and Music},
organization = {MML 2020},
keywords = {GRE19-04, HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Castellanos, F. J.; Gallego, A. J.; Calvo-Zaragoza, J.
Automatic scale estimation for music score images Journal Article
In: Expert Systems with Applications, vol. 158, no. 113590, 2020, ISSN: 0957-4174.
BibTeX | Tags:
@article{k450,
title = {Automatic scale estimation for music score images},
author = {F. J. Castellanos and A. J. Gallego and J. Calvo-Zaragoza},
issn = {0957-4174},
year = {2020},
date = {2020-06-01},
urldate = {2020-06-01},
journal = {Expert Systems with Applications},
volume = {158},
number = {113590},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Vayá, Saborit M. Iglesia; Montell, J. A.; Pertusa, A.; Bustos, A.; Cazorla, M.; Galant, J.; Barber, X.; Orozco-Beltrán, D.; García-García, F.; Caparrós, M.; González, G.; Salinas, J. M.
BIMCV COVID-19+: a large annotated dataset of RX and CT images from COVID-19 patients Journal Article
In: ArXiv:2006.01174, 2020.
@article{k453,
title = {BIMCV COVID-19+: a large annotated dataset of RX and CT images from COVID-19 patients},
author = {Saborit M. Iglesia Vayá and J. A. Montell and A. Pertusa and A. Bustos and M. Cazorla and J. Galant and X. Barber and D. Orozco-Beltrán and F. García-García and M. Caparrós and G. González and J. M. Salinas},
year = {2020},
date = {2020-06-01},
journal = {ArXiv:2006.01174},
abstract = {This paper describes BIMCV COVID-19+, a large dataset from the Valencian Region Medical ImageBank (BIMCV) containing chest X-ray images CXR (CR, DX) and computed tomography (CT) imaging of COVID-19+ patients along with their radiological findings and locations, pathologies, radiological reports (in Spanish), DICOM metadata, Polymerase chain reaction (PCR), Immunoglobulin G (IgG) and Immunoglobulin M (IgM) diagnostic antibody tests. The findings have been mapped onto standard Unified Medical Language System (UMLS) terminology and cover a wide spectrum of thoracic entities, unlike the considerably more reduced number of entities annotated in previous datasets. Images are stored in high resolution and entities are localized with anatomical labels and stored in a Medical Imaging Data Structure (MIDS) format. In addition, 10 images were annotated by a team of radiologists to include semantic segmentation of radiological findings. This first iteration of the database includes 1,380 CX, 885 DX and 163 CT studies from 1,311 COVID-19+ patients. This is, to the best of our knowledge, the largest COVID-19+ dataset of images available in an open format. The dataset can be downloaded from this http URL.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
This paper describes BIMCV COVID-19+, a large dataset from the Valencian Region Medical ImageBank (BIMCV) containing chest X-ray images CXR (CR, DX) and computed tomography (CT) imaging of COVID-19+ patients along with their radiological findings and locations, pathologies, radiological reports (in Spanish), DICOM metadata, Polymerase chain reaction (PCR), Immunoglobulin G (IgG) and Immunoglobulin M (IgM) diagnostic antibody tests. The findings have been mapped onto standard Unified Medical Language System (UMLS) terminology and cover a wide spectrum of thoracic entities, unlike the considerably more reduced number of entities annotated in previous datasets. Images are stored in high resolution and entities are localized with anatomical labels and stored in a Medical Imaging Data Structure (MIDS) format. In addition, 10 images were annotated by a team of radiologists to include semantic segmentation of radiological findings. This first iteration of the database includes 1,380 CX, 885 DX and 163 CT studies from 1,311 COVID-19+ patients. This is, to the best of our knowledge, the largest COVID-19+ dataset of images available in an open format. The dataset can be downloaded from this http URL.
González, G.; Bustos, A.; Salinas, J. M.; Iglesia-Vaya, M.; Galant, J.; Cano-Espinosa, C.; Barber, X.; Orozco-Beltrán, D.; Cazorla, M.; Pertusa, A.
UMLS-ChestNet: A deep convolutional neural network for radiological findings, differential diagnoses and localizations of COVID-19 in chest x-rays Journal Article
In: arxiv:2006.05274, 2020.
@article{k454,
title = {UMLS-ChestNet: A deep convolutional neural network for radiological findings, differential diagnoses and localizations of COVID-19 in chest x-rays},
author = {G. González and A. Bustos and J. M. Salinas and M. Iglesia-Vaya and J. Galant and C. Cano-Espinosa and X. Barber and D. Orozco-Beltrán and M. Cazorla and A. Pertusa},
year = {2020},
date = {2020-06-01},
journal = {arxiv:2006.05274},
abstract = {In this work we present a method for the detection of radiological findings, their location and differential diagnoses from chest x-rays. Unlike prior works that focus on the detection of few pathologies, we use a hierarchical taxonomy mapped to the Unified Medical Language System (UMLS) terminology to identify 189 radiological findings, 22 differential diagnosis and 122 anatomic locations, including ground glass opacities, infiltrates, consolidations and other radiological findings compatible with COVID-19. We train the system on one large database of 92,594 frontal chest x-rays (AP or PA, standing, supine or decubitus) and a second database of 2,065 frontal images of COVID-19 patients identified by at least one positive Polymerase Chain Reaction (PCR) test. The reference labels are obtained through natural language processing of the radiological reports. On 23,159 test images, the proposed neural network obtains an AUC of 0.94 for the diagnosis of COVID-19. To our knowledge, this work uses the largest chest x-ray dataset of COVID-19 positive cases to date and is the first one to use a hierarchical labeling schema and to provide interpretability of the results, not only by using network attention methods, but also by indicating the radiological findings that have led to the diagnosis.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
In this work we present a method for the detection of radiological findings, their location and differential diagnoses from chest x-rays. Unlike prior works that focus on the detection of few pathologies, we use a hierarchical taxonomy mapped to the Unified Medical Language System (UMLS) terminology to identify 189 radiological findings, 22 differential diagnosis and 122 anatomic locations, including ground glass opacities, infiltrates, consolidations and other radiological findings compatible with COVID-19. We train the system on one large database of 92,594 frontal chest x-rays (AP or PA, standing, supine or decubitus) and a second database of 2,065 frontal images of COVID-19 patients identified by at least one positive Polymerase Chain Reaction (PCR) test. The reference labels are obtained through natural language processing of the radiological reports. On 23,159 test images, the proposed neural network obtains an AUC of 0.94 for the diagnosis of COVID-19. To our knowledge, this work uses the largest chest x-ray dataset of COVID-19 positive cases to date and is the first one to use a hierarchical labeling schema and to provide interpretability of the results, not only by using network attention methods, but also by indicating the radiological findings that have led to the diagnosis.
Thomae, M. E.; Ríos-Vila, A.; Calvo-Zaragoza, J.; Rizo, D.; Iñesta, J. M.
Retrieving Music Semantics from Optical Music Recognition by Machine Translation Proceedings Article
In: Luca, Elsa De; Flanders, Julia (Ed.): Music Encoding Conference Proceedings 26-29 May, 2020 Tufts University, Boston (USA), pp. 19-24, Music Encoding Initiative 2020.
Links | BibTeX | Tags: HispaMus
@inproceedings{k446,
title = {Retrieving Music Semantics from Optical Music Recognition by Machine Translation},
author = {M. E. Thomae and A. Ríos-Vila and J. Calvo-Zaragoza and D. Rizo and J. M. Iñesta},
editor = {Elsa De Luca and Julia Flanders},
url = {https://grfia.dlsi.ua.es/repositori/grfia/pubs/446/music_encoding_conference_proceedings_2020.pdf},
year = {2020},
date = {2020-05-01},
booktitle = {Music Encoding Conference Proceedings 26-29 May, 2020 Tufts University, Boston (USA)},
pages = {19-24},
organization = {Music Encoding Initiative},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Martinez-Martin, E.; Pertusa, A.; Costa, A.; Gomez-Donoso, F.; Escalona, F.; Viejo, D.; Cazorla, M.
The competition as assessment tool in robotics engineering Proceedings Article
In: INTED2020 Proceedings, pp. 1082-1085, Valencia, Spain, 2020, ISBN: 978-84-09-17939-8.
BibTeX | Tags:
@inproceedings{k460,
title = {The competition as assessment tool in robotics engineering},
author = {E. Martinez-Martin and A. Pertusa and A. Costa and F. Gomez-Donoso and F. Escalona and D. Viejo and M. Cazorla},
isbn = {978-84-09-17939-8},
year = {2020},
date = {2020-03-01},
booktitle = {INTED2020 Proceedings},
pages = {1082-1085},
address = {Valencia, Spain},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Calvo-Zaragoza, J.; Jr., J. Hajic; Pacha, A.
Understanding Optical Music Recognition Journal Article
In: ACM Computing Surveys, vol. 53, no. 4, pp. 77, 2020.
@article{k426,
title = {Understanding Optical Music Recognition},
author = {J. Calvo-Zaragoza and J. Hajic Jr. and A. Pacha},
year = {2020},
date = {2020-01-01},
journal = {ACM Computing Surveys},
volume = {53},
number = {4},
pages = {77},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {article}
}
Cuevas-Velasquez, H.; Gallego, A. J.; Fisher, R. B.
Segmentation and 3D reconstruction of rose plants from stereoscopic images Journal Article
In: Computers and Electronics in Agriculture, vol. 171, pp. 105296, 2020, ISSN: 0168-1699.
@article{k440,
title = {Segmentation and 3D reconstruction of rose plants from stereoscopic images},
author = {H. Cuevas-Velasquez and A. J. Gallego and R. B. Fisher},
issn = {0168-1699},
year = {2020},
date = {2020-01-01},
urldate = {2020-01-01},
journal = {Computers and Electronics in Agriculture},
volume = {171},
pages = {105296},
abstract = {The method proposed in this paper is part of the vision module of a garden robot capable of navigating towards rose bushes and clip them according to a set of pruning rules. The method is responsible for performing the segmentation of the branches and recovering their morphology in 3D. The obtained reconstruction allows the manipulator of the robot to select the candidate branches to be pruned. This method first obtains a stereo pair of images and calculates the disparity image using block matching and the segmentation of the branches using a Fully Convolutional Neuronal Network modified to return a map with the probability at the pixel level of the presence of a branch. A post-processing step combines the segmentation and the disparity in order to improve the results. Then, the skeleton of the plant and the branching structure are calculated, and finally, the 3D reconstruction is obtained. The proposed approach is evaluated with five different datasets, three of them compiled by the authors and two from the state of the art, including indoor and outdoor scenes with uncontrolled environments. The different steps of the proposed pipeline are evaluated and compared with other state-of-the-art methods, showing that the accuracy of the segmentation improves other methods for this task, even with variable lighting, and also that the skeletonization and the reconstruction processes obtain robust results.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The method proposed in this paper is part of the vision module of a garden robot capable of navigating towards rose bushes and clip them according to a set of pruning rules. The method is responsible for performing the segmentation of the branches and recovering their morphology in 3D. The obtained reconstruction allows the manipulator of the robot to select the candidate branches to be pruned. This method first obtains a stereo pair of images and calculates the disparity image using block matching and the segmentation of the branches using a Fully Convolutional Neuronal Network modified to return a map with the probability at the pixel level of the presence of a branch. A post-processing step combines the segmentation and the disparity in order to improve the results. Then, the skeleton of the plant and the branching structure are calculated, and finally, the 3D reconstruction is obtained. The proposed approach is evaluated with five different datasets, three of them compiled by the authors and two from the state of the art, including indoor and outdoor scenes with uncontrolled environments. The different steps of the proposed pipeline are evaluated and compared with other state-of-the-art methods, showing that the accuracy of the segmentation improves other methods for this task, even with variable lighting, and also that the skeletonization and the reconstruction processes obtain robust results.
Cuevas-Velasquez, H.; Gallego, A. J.; Tylecek, R.; Hemming, J.; Tuijl, B.; Mencarelli, A.; Fisher, R. B.
Real-time Stereo Visual Servoing for Rose Pruning with Robotic Arm Proceedings Article
In: International Conference on Robotics and Automation (ICRA), IEEE, Paris, France, 2020.
@inproceedings{k441,
title = {Real-time Stereo Visual Servoing for Rose Pruning with Robotic Arm},
author = {H. Cuevas-Velasquez and A. J. Gallego and R. Tylecek and J. Hemming and B. Tuijl and A. Mencarelli and R. B. Fisher},
year = {2020},
date = {2020-01-01},
urldate = {2020-01-01},
booktitle = {International Conference on Robotics and Automation (ICRA)},
publisher = {IEEE},
address = {Paris, France},
abstract = {The paper presents a working pipeline which integrates hardware and software in an automated robotic rose cutter. To the best of our knowledge, this is the first robot able to prune rose bushes in a natural environment. Unlike similar approaches like tree stem cutting, the proposed method does not require to scan the full plant, have multiple cameras around the bush, or assume that a stem does not move. It relies on a single stereo camera mounted on the end-effector of the robot and real-time visual servoing to navigate to the desired cutting location on the stem. The evaluation of the whole pipeline shows a good performance in a garden with unconstrained conditions, where finding and approaching a specific location on a stem is challenging due to occlusions caused by other stems and dynamic changes caused by the wind.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The paper presents a working pipeline which integrates hardware and software in an automated robotic rose cutter. To the best of our knowledge, this is the first robot able to prune rose bushes in a natural environment. Unlike similar approaches like tree stem cutting, the proposed method does not require to scan the full plant, have multiple cameras around the bush, or assume that a stem does not move. It relies on a single stereo camera mounted on the end-effector of the robot and real-time visual servoing to navigate to the desired cutting location on the stem. The evaluation of the whole pipeline shows a good performance in a garden with unconstrained conditions, where finding and approaching a specific location on a stem is challenging due to occlusions caused by other stems and dynamic changes caused by the wind.
Oncina, J.; Iñesta, J. M.; Micó, L.
Adaptively Learning to Recognize Symbols in Handwritten Early Music Book Chapter
In: Cellier, Driessens Peggy (Ed.): Machine Learning and Knowledge Discovery in Databases, Chapter 40, pp. 470–477, Springer, 2020, ISBN: 978-3-030-43887-6.
Abstract | BibTeX | Tags: HispaMus
@inbook{k443,
title = {Adaptively Learning to Recognize Symbols in Handwritten Early Music},
author = {J. Oncina and J. M. Iñesta and L. Micó},
editor = {Driessens Peggy Cellier},
isbn = {978-3-030-43887-6},
year = {2020},
date = {2020-01-01},
urldate = {2020-01-01},
booktitle = {Machine Learning and Knowledge Discovery in Databases},
pages = {470--477},
publisher = {Springer},
chapter = {40},
abstract = {Human supervision is necessary for a correct edition and publication of handwritten early music collections. The output of an optical music recognition system for that kind of documents may contain a significant number of errors, making it tedious to correct for a human expert. An adequate strategy is needed to optimize the human feedback information during the correction stage to adapt the classifier to the specificities of each manuscript. In this paper, we compare the performance of a neural system, difficult and slow to be retrained, and a nearest neighbor strategy, based on the neural codes provided by a neural net, trained offline, used as a feature extractor.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inbook}
}
Human supervision is necessary for a correct edition and publication of handwritten early music collections. The output of an optical music recognition system for that kind of documents may contain a significant number of errors, making it tedious to correct for a human expert. An adequate strategy is needed to optimize the human feedback information during the correction stage to adapt the classifier to the specificities of each manuscript. In this paper, we compare the performance of a neural system, difficult and slow to be retrained, and a nearest neighbor strategy, based on the neural codes provided by a neural net, trained offline, used as a feature extractor.
Gallego, A. J.; Calvo-Zaragoza, J.; Rico-Juan, J. R.
Insights Into Efficient k-Nearest Neighbor Classification With Convolutional Neural Codes Journal Article
In: IEEE Access, vol. 8, pp. 99312-99326, 2020, ISSN: 2169-3536.
Abstract | BibTeX | Tags: HispaMus
@article{k444,
title = {Insights Into Efficient k-Nearest Neighbor Classification With Convolutional Neural Codes},
author = {A. J. Gallego and J. Calvo-Zaragoza and J. R. Rico-Juan},
issn = {2169-3536},
year = {2020},
date = {2020-01-01},
urldate = {2020-01-01},
journal = {IEEE Access},
volume = {8},
pages = {99312-99326},
abstract = {The increasing consideration of Convolutional Neural Networks (CNN) has not prevented the use of the k-Nearest Neighbor (kNN) method. In fact, a hybrid CNN-kNN approach is an interesting option in which the network specializes in feature extraction through its activations (Neural Codes), while the kNN has the advantage of performing a retrieval by means of similarity. However, this hybrid approach also has the disadvantages of the kNN search, and especially its high computational cost which is, in principle, undesirable for large-scale data. In this paper, we present the first comprehensive study of efficient kNN search algorithms using this hybrid CNN-kNN approach. This has been done by considering up to 16 different algorithms, each of which is evaluated with a different parametrization, in 7 datasets of heterogeneous composition. Our results show that no single algorithm is capable of covering all aspects, but rather that each family of algorithms is better suited to specific aspects of the problem. This signifies that Fast Similarity Search algorithms maintain their performance, but do not reduce the cost as much as the Data Reduction family does. In turn, the Approximated Similarity Search family is postulated as a good option when attempting to balance accuracy and efficiency. The experiments also suggest that considering statistical transformation algorithms such as Linear Discriminant Analysis might be useful in certain cases.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {article}
}
The increasing consideration of Convolutional Neural Networks (CNN) has not prevented the use of the k-Nearest Neighbor (kNN) method. In fact, a hybrid CNN-kNN approach is an interesting option in which the network specializes in feature extraction through its activations (Neural Codes), while the kNN has the advantage of performing a retrieval by means of similarity. However, this hybrid approach also has the disadvantages of the kNN search, and especially its high computational cost which is, in principle, undesirable for large-scale data. In this paper, we present the first comprehensive study of efficient kNN search algorithms using this hybrid CNN-kNN approach. This has been done by considering up to 16 different algorithms, each of which is evaluated with a different parametrization, in 7 datasets of heterogeneous composition. Our results show that no single algorithm is capable of covering all aspects, but rather that each family of algorithms is better suited to specific aspects of the problem. This signifies that Fast Similarity Search algorithms maintain their performance, but do not reduce the cost as much as the Data Reduction family does. In turn, the Approximated Similarity Search family is postulated as a good option when attempting to balance accuracy and efficiency. The experiments also suggest that considering statistical transformation algorithms such as Linear Discriminant Analysis might be useful in certain cases.
Valero-Mas, J. J.; Castellanos, F. J.
Data Reduction in the String Space for Efficient kNN Classification through Space Partitioning Journal Article
In: Applied Sciences, vol. 10, no. 10, pp. 3356, 2020.
@article{k445,
title = {Data Reduction in the String Space for Efficient kNN Classification through Space Partitioning},
author = {J. J. Valero-Mas and F. J. Castellanos},
year = {2020},
date = {2020-01-01},
journal = {Applied Sciences},
volume = {10},
number = {10},
pages = {3356},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {article}
}
Pertusa, A.; Román, M. A.
Data representations for audio-to-score monophonic music transcription Journal Article
In: Expert Systems with Applications, vol. 162, pp. 113769, 2020, ISSN: 0957-4174.
@article{k447,
title = {Data representations for audio-to-score monophonic music transcription},
author = {A. Pertusa and M. A. Román},
issn = {0957-4174},
year = {2020},
date = {2020-01-01},
urldate = {2020-01-01},
journal = {Expert Systems with Applications},
volume = {162},
pages = {113769},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {article}
}
Ríos-Vila, A.; Calvo-Zaragoza, J.; Iñesta, J. M.
Exploring the two-dimensional nature of music notation for score recognition with end-to-end approaches Proceedings Article
In: Proc. of 17th International Conference on Frontiers in Handwriting Recognition (ICFHR), pp. 193–198, IEEE Computer Society IEEE, Dormund (Germany), 2020, ISBN: 978-1-7281-9966-5.
Abstract | BibTeX | Tags: HispaMus
@inproceedings{k461,
title = {Exploring the two-dimensional nature of music notation for score recognition with end-to-end approaches},
author = {A. Ríos-Vila and J. Calvo-Zaragoza and J. M. Iñesta},
isbn = {978-1-7281-9966-5},
year = {2020},
date = {2020-01-01},
booktitle = {Proc. of 17th International Conference on Frontiers in Handwriting Recognition (ICFHR)},
pages = {193--198},
publisher = {IEEE},
address = {Dormund (Germany)},
organization = {IEEE Computer Society},
abstract = {Optical Music Recognition workflows perform several steps to retrieve the content in music score images, being symbol recognition one of the key stages. State-of-the-art approaches for this stage currently address the coding of the output symbols as if they were plain text characters. However, music symbols have a two-dimensional nature that is ignored in these approaches. In this paper, we explore alternative output representations to perform music symbol recognition with state-of-the-art end-to-end neural technologies. We propose and describe new output representations which take into account the mentioned two-dimensional nature. We seek answers to the question of whether it is possible to obtain better recognition results in both printed and handwritten music scores. In this analysis, we compare the results given using three output encodings and two neural approaches. We found that one of the proposed encodings outperforms the results obtained by the standard one. This permits us to conclude that it is interesting to keep researching on this topic to improve end-to-end music score recognition.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Optical Music Recognition workflows perform several steps to retrieve the content in music score images, being symbol recognition one of the key stages. State-of-the-art approaches for this stage currently address the coding of the output symbols as if they were plain text characters. However, music symbols have a two-dimensional nature that is ignored in these approaches. In this paper, we explore alternative output representations to perform music symbol recognition with state-of-the-art end-to-end neural technologies. We propose and describe new output representations which take into account the mentioned two-dimensional nature. We seek answers to the question of whether it is possible to obtain better recognition results in both printed and handwritten music scores. In this analysis, we compare the results given using three output encodings and two neural approaches. We found that one of the proposed encodings outperforms the results obtained by the standard one. This permits us to conclude that it is interesting to keep researching on this topic to improve end-to-end music score recognition.
Rizo, D.; Calvo-Zaragoza, J.; Iñesta, J. M.
Preservación del patrimonio musical mediante transcripción asistida por ordenador Book Chapter
In: Isusi-Fagoaga, R.; Serrano, F. Villanueva (Ed.): La música de la Corona d’Aragó: investigació, transferència i educació, Chapter 15, pp. 275–298, Universitat de València, 2020, ISBN: 978-84-09-19985-3.
@inbook{k466,
title = {Preservación del patrimonio musical mediante transcripción asistida por ordenador},
author = {D. Rizo and J. Calvo-Zaragoza and J. M. Iñesta},
editor = {R. Isusi-Fagoaga and F. Villanueva Serrano},
isbn = {978-84-09-19985-3},
year = {2020},
date = {2020-01-01},
urldate = {2020-01-01},
booktitle = {La música de la Corona d’Aragó: investigació, transferència i educació},
pages = {275--298},
publisher = {Universitat de València},
chapter = {15},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inbook}
}
Ortega-Bastida, J.; Gallego, A. J.; Rico-Juan, J. R.; Albarrán, P.
Regional gross domestic product prediction using Twitter deep learning representations Proceedings Article
In: 17th IADIS International Conference on Applied Computing (IADIS AC), pp. 89–96, Lisboa, Portugal, 2020, ISBN: 978-989-8704-24-5.
BibTeX | Tags:
@inproceedings{k472,
title = {Regional gross domestic product prediction using Twitter deep learning representations},
author = {J. Ortega-Bastida and A. J. Gallego and J. R. Rico-Juan and P. Albarrán},
isbn = {978-989-8704-24-5},
year = {2020},
date = {2020-01-01},
urldate = {2020-01-01},
booktitle = {17th IADIS International Conference on Applied Computing (IADIS AC)},
pages = {89--96},
address = {Lisboa, Portugal},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
2019
Calvo-Zaragoza, J.; Toselli, A. H.; Vidal, E.
Handwritten Music Recognition for Mensural Notation with Convolutional Recurrent Neural Networks Journal Article
In: Pattern Recognition Letters, vol. 128, pp. 115–121, 2019.
Links | BibTeX | Tags: HispaMus
@article{k415,
title = {Handwritten Music Recognition for Mensural Notation with Convolutional Recurrent Neural Networks},
author = {J. Calvo-Zaragoza and A. H. Toselli and E. Vidal},
url = {https://grfia.dlsi.ua.es/repositori/grfia/pubs/415/1-s2.0-S0167865519302338-main.pdf},
year = {2019},
date = {2019-12-01},
journal = {Pattern Recognition Letters},
volume = {128},
pages = {115--121},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {article}
}
Iñesta, J. M.; Rizo, D.; Calvo-Zaragoza, J.
MuRET as a software for the transcription of historical archives Proceedings Article
In: Proceedings of the 2nd Workshop on Reading Music Systems, WoRMS, pp. 12–15, Delft (The Nederlands), 2019.
Abstract | Links | BibTeX | Tags: HispaMus
@inproceedings{k437,
title = {MuRET as a software for the transcription of historical archives},
author = {J. M. Iñesta and D. Rizo and J. Calvo-Zaragoza},
url = {https://grfia.dlsi.ua.es/repositori/grfia/pubs/437/worms-2019-inesta.pdf},
year = {2019},
date = {2019-11-01},
booktitle = {Proceedings of the 2nd Workshop on Reading Music Systems, WoRMS},
pages = {12--15},
address = {Delft (The Nederlands)},
abstract = {The transcription process from historical hand- written music manuscripts to a structured digital encoding has been traditionally performed following a fully manual workflow. At most, it has received some technological support in particular stages, like optical music recognition (OMR) of the source images, or transcription to modern notation with music edition applications. Currently, there is no mature and stable enough solution for the OMR problem, and the most used music editors do not support early notations, such as the mensural notation. A new tool called MUsic Recognition, Encoding, and Transcription (MuRET) has been developed, which covers all transcription phases, from the manuscript image to the encoded digital content. MuRET is designed as a machine-learning based research tool, allowing different processing approaches to be used, and producing both the expected transcribed contents in standard encodings and data for the study of the transcription process.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
The transcription process from historical hand- written music manuscripts to a structured digital encoding has been traditionally performed following a fully manual workflow. At most, it has received some technological support in particular stages, like optical music recognition (OMR) of the source images, or transcription to modern notation with music edition applications. Currently, there is no mature and stable enough solution for the OMR problem, and the most used music editors do not support early notations, such as the mensural notation. A new tool called MUsic Recognition, Encoding, and Transcription (MuRET) has been developed, which covers all transcription phases, from the manuscript image to the encoded digital content. MuRET is designed as a machine-learning based research tool, allowing different processing approaches to be used, and producing both the expected transcribed contents in standard encodings and data for the study of the transcription process.
Rizo, D.
DLfM '19: 6th International Conference on Digital Libraries for Musicology Book
ACM, The Hague, Netherlands, 2019, ISBN: 978-1-4503-7239-8.
@book{k438,
title = {DLfM '19: 6th International Conference on Digital Libraries for Musicology},
author = {D. Rizo},
editor = {D. Rizo},
isbn = {978-1-4503-7239-8},
year = {2019},
date = {2019-11-01},
urldate = {2019-11-01},
publisher = {ACM},
address = {The Hague, Netherlands},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {book}
}
Ríos-Vila, A.; Calvo-Zaragoza, J.; Rizo, D.; Iñesta, J. M.
ReadSco: An Open-Source Web-Based Optical Music Recognition Tool Proceedings Article
In: Calvo-Zaragoza, J.; Pacha, A. (Ed.): Proc. of the 2nd International Workshop on Reading Music Systems, WoRMS 2019, pp. 27-30, Delft (The Netherlands), 2019.
Abstract | Links | BibTeX | Tags: HispaMus
@inproceedings{k439,
title = {ReadSco: An Open-Source Web-Based Optical Music Recognition Tool},
author = {A. Ríos-Vila and J. Calvo-Zaragoza and D. Rizo and J. M. Iñesta},
editor = {J. Calvo-Zaragoza and A. Pacha},
url = {https://grfia.dlsi.ua.es/repositori/grfia/pubs/439/ReadSco_Worms2019.pdf},
year = {2019},
date = {2019-11-01},
urldate = {2019-11-01},
booktitle = {Proc. of the 2nd International Workshop on Reading Music Systems, WoRMS 2019},
pages = {27-30},
address = {Delft (The Netherlands)},
abstract = {We introduce READSCO, an open-source web-based community tool for Optical Music Recognition. READSCO aims to serve as a connection between research results and practical use. We describe the design decisions considered to both favours a rapid integration of new advances from the research field and
facilitate the community’s participation in its development. The project is still in its planning phase, so this work is a good opportunity to present the main idea and get direct feedback from other researchers.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
We introduce READSCO, an open-source web-based community tool for Optical Music Recognition. READSCO aims to serve as a connection between research results and practical use. We describe the design decisions considered to both favours a rapid integration of new advances from the research field and
facilitate the community’s participation in its development. The project is still in its planning phase, so this work is a good opportunity to present the main idea and get direct feedback from other researchers.
facilitate the community’s participation in its development. The project is still in its planning phase, so this work is a good opportunity to present the main idea and get direct feedback from other researchers.
Iñesta, J. M.; Rizo, D.; Calvo-Zaragoza, J.
Recognition, encoding, and transcription of early mensural handwritten music Proceedings Article
In: procedings of the 13th International Workshop on Graphics Recognition (GREC 2019), pp. 13-14, Sydney (Australia), 2019.
@inproceedings{k416,
title = {Recognition, encoding, and transcription of early mensural handwritten music},
author = {J. M. Iñesta and D. Rizo and J. Calvo-Zaragoza},
year = {2019},
date = {2019-09-01},
booktitle = {procedings of the 13th International Workshop on Graphics Recognition (GREC 2019)},
pages = {13-14},
address = {Sydney (Australia)},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Micó, L.; Oncina, J.; Iñesta, J. M.
Adaptively learning to recognize symbols in handwritten early music Book Chapter
In: Conklin, D. (Ed.): procedings of the ECML/PKDD 12th International Workshop on Machine Learning and Music (MML 2019), Chapter (to appear), Springer, Würzburg (Germany), 2019.
Abstract | Links | BibTeX | Tags: HispaMus
@inbook{k417,
title = {Adaptively learning to recognize symbols in handwritten early music},
author = {L. Micó and J. Oncina and J. M. Iñesta},
editor = {D. Conklin},
url = {https://grfia.dlsi.ua.es/repositori/grfia/pubs/417/Learning_to_recognize_handwritten_early_music.pdf},
year = {2019},
date = {2019-09-01},
booktitle = {procedings of the ECML/PKDD 12th International Workshop on Machine Learning and Music (MML 2019)},
publisher = {Springer},
address = {Würzburg (Germany)},
chapter = {(to appear)},
abstract = {Human supervision is necessary for a correct edition and publication of handwritten early music collections. The output of an op- tical music recognition system for that kind of documents may contain a significant number of errors, making it tedious to correct for a human expert. An adequate strategy is needed to optimize the human feedback information during the correction stage to adapt the classifier to the specificities of each manuscript. In this paper, we compare the perfor- mance of a neural system, difficult and slow to be retrained, and a nearest neighbor strategy, based on the neural codes provided by a neural net, trained offline, used as a feature extractor.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inbook}
}
Human supervision is necessary for a correct edition and publication of handwritten early music collections. The output of an op- tical music recognition system for that kind of documents may contain a significant number of errors, making it tedious to correct for a human expert. An adequate strategy is needed to optimize the human feedback information during the correction stage to adapt the classifier to the specificities of each manuscript. In this paper, we compare the perfor- mance of a neural system, difficult and slow to be retrained, and a nearest neighbor strategy, based on the neural codes provided by a neural net, trained offline, used as a feature extractor.
Alfaro-Contreras, M.; Calvo-Zaragoza, J.; Iñesta, J. M.
Approaching End-to-end Optical Music Recognition for Homophonic Scores Proceedings Article
In: Iberian Conference on Pattern Recognition and Image Analysis, IbPRIA 2019, pp. 147–158, Springer, Madrid, 2019, ISBN: 978-3-030-31321-0.
Links | BibTeX | Tags: HispaMus
@inproceedings{k419,
title = {Approaching End-to-end Optical Music Recognition for Homophonic Scores},
author = {M. Alfaro-Contreras and J. Calvo-Zaragoza and J. M. Iñesta},
url = {https://link.springer.com/chapter/10.1007/978-3-030-31321-0_13},
isbn = {978-3-030-31321-0},
year = {2019},
date = {2019-08-01},
urldate = {2019-08-01},
booktitle = {Iberian Conference on Pattern Recognition and Image Analysis, IbPRIA 2019},
pages = {147--158},
publisher = {Springer},
address = {Madrid},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Mateiu, T. N.; Gallego, A. J.; Calvo-Zaragoza, J.
Domain Adaptation for Handwritten Symbol Recognition: A Case of Study in Old Music Manuscripts Proceedings Article
In: Iberian Conference on Pattern Recognition and Image Analysis, IbPRIA, pp. 135–146, Springer International Publishing, Madrid, Spain, 2019, ISBN: 978-3-030-31321-0.
Abstract | BibTeX | Tags: HispaMus
@inproceedings{k410,
title = {Domain Adaptation for Handwritten Symbol Recognition: A Case of Study in Old Music Manuscripts},
author = {T. N. Mateiu and A. J. Gallego and J. Calvo-Zaragoza},
isbn = {978-3-030-31321-0},
year = {2019},
date = {2019-07-01},
urldate = {2019-07-01},
booktitle = {Iberian Conference on Pattern Recognition and Image Analysis, IbPRIA},
pages = {135--146},
publisher = {Springer International Publishing},
address = {Madrid, Spain},
abstract = {The existence of a large amount of untranscripted music manuscripts has caused initiatives that use Machine Learning (ML) for Optical Music Recognition, in order to efficiently transcribe the music sources into a machine-readable format. Although most music manuscript are similar in nature, they inevitably vary from one another. This fact can negatively influence the complexity of the classification task because most ML models fail to transfer their knowledge from one domain to another, thereby requiring learning from scratch on new domains after manually labeling new data. This work studies the ability of a Domain Adversarial Neural Network for domain adaptation in the context of classifying handwritten music symbols. The main idea is to exploit the knowledge of a specific manuscript to classify symbols from different (unlabeled) manuscripts. The reported results are promising, obtaining a substantial improvement over a conventional Convolutional Neural Network approach, which can be used as a basis for future research.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
The existence of a large amount of untranscripted music manuscripts has caused initiatives that use Machine Learning (ML) for Optical Music Recognition, in order to efficiently transcribe the music sources into a machine-readable format. Although most music manuscript are similar in nature, they inevitably vary from one another. This fact can negatively influence the complexity of the classification task because most ML models fail to transfer their knowledge from one domain to another, thereby requiring learning from scratch on new domains after manually labeling new data. This work studies the ability of a Domain Adversarial Neural Network for domain adaptation in the context of classifying handwritten music symbols. The main idea is to exploit the knowledge of a specific manuscript to classify symbols from different (unlabeled) manuscripts. The reported results are promising, obtaining a substantial improvement over a conventional Convolutional Neural Network approach, which can be used as a basis for future research.
Gallego, A. J.; Pertusa, A.; Bernabeu, M.
Multi-Label Logo Classification using Convolutional Neural Networks Proceedings Article
In: Iberian Conference on Pattern Recognition and Image Analysis, IbPRIA, pp. 485–497, Springer International Publishing, Madrid, Spain, 2019, ISBN: 978-3-030-31332-6.
Abstract | BibTeX | Tags: HispaMus
@inproceedings{k411,
title = {Multi-Label Logo Classification using Convolutional Neural Networks},
author = {A. J. Gallego and A. Pertusa and M. Bernabeu},
isbn = {978-3-030-31332-6},
year = {2019},
date = {2019-07-01},
urldate = {2019-07-01},
booktitle = {Iberian Conference on Pattern Recognition and Image Analysis, IbPRIA},
pages = {485--497},
publisher = {Springer International Publishing},
address = {Madrid, Spain},
abstract = {The classification of logos is a particular case within computer vision since they have their own characteristics. Logos can contain only text, iconic images or a combination of both, and they usually include figurative symbols designed by experts that vary substantially besides they may share the same semantics. This work presents a method for multi-label classification and retrieval of logo images. For this, Convolutional Neural Networks (CNN) are trained to classify logos from the European Union TradeMark (EUTM) dataset according to their colors, shapes, sectors and figurative designs. An auto-encoder is also trained to learn representations of the input images. Once trained, the neural codes from the last convolutional layers in the CNN and the central layer of the auto-encoder can be used to perform similarity search through kNN, allowing us to obtain the most similar logos based on their color, shape, sector, figurative elements, overall features, or a weighted combination of them provided by the user. To the best of our knowledge, this is the first multi-label classification method for logos, and the only one that allows retrieving a ranking of images with these criteria provided by the user.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
The classification of logos is a particular case within computer vision since they have their own characteristics. Logos can contain only text, iconic images or a combination of both, and they usually include figurative symbols designed by experts that vary substantially besides they may share the same semantics. This work presents a method for multi-label classification and retrieval of logo images. For this, Convolutional Neural Networks (CNN) are trained to classify logos from the European Union TradeMark (EUTM) dataset according to their colors, shapes, sectors and figurative designs. An auto-encoder is also trained to learn representations of the input images. Once trained, the neural codes from the last convolutional layers in the CNN and the central layer of the auto-encoder can be used to perform similarity search through kNN, allowing us to obtain the most similar logos based on their color, shape, sector, figurative elements, overall features, or a weighted combination of them provided by the user. To the best of our knowledge, this is the first multi-label classification method for logos, and the only one that allows retrieving a ranking of images with these criteria provided by the user.
Nuñez-Alcover, Alicia; León, P. J. Ponce; Calvo-Zaragoza, J.
Glyph and Position Classification of Music Symbols in Early Music Manuscripts Proceedings Article
In: Morales, A.; Fierrez, J.; Sánchez, J. Salvador; Ribeiro, B. (Ed.): Proc. of the 9th Iberian Conference on Pattern Recognition and Image Analysis, LNCS vol. 11867, pp. 159-168, AERFAI, APRP Springer, Madrid, Spain, 2019, ISBN: 978-3-030-31331-9.
Abstract | Links | BibTeX | Tags: HispaMus
@inproceedings{k420,
title = {Glyph and Position Classification of Music Symbols in Early Music Manuscripts},
author = {Alicia Nuñez-Alcover and P. J. Ponce León and J. Calvo-Zaragoza},
editor = {A. Morales and J. Fierrez and J. Salvador Sánchez and B. Ribeiro},
url = {https://grfia.dlsi.ua.es/repositori/grfia/pubs/420/IbPRIA2019_paper_91-1.pdf},
isbn = {978-3-030-31331-9},
year = {2019},
date = {2019-07-01},
booktitle = {Proc. of the 9th Iberian Conference on Pattern Recognition and Image Analysis, LNCS vol. 11867},
pages = {159-168},
publisher = {Springer},
address = {Madrid, Spain},
organization = {AERFAI, APRP},
abstract = {Optical Music Recognition is a field of research that automates the reading of musical scores so as to transcribe their content into a structured digital format. When dealing with music manuscripts, the traditional workflow establishes separate stages of detection and classification of musical symbols. In the latter, most of the research has focused on detecting musical glyphs, ignoring that the meaning of a musical symbol is defined by two components: its glyph and its position within the staff. In this paper we study how to perform both glyph and position classification of handwritten musical symbols in early music manuscripts written in white Mensural notation, a common notation system used for the most part of the XVI and XVII centuries. We make use of Convolutional Neural Networks as the classification method, and we tested several alternatives such as using independent models for each component, combining label spaces, or using both multi-input and multi-output models. Our results on early music manuscripts provide insights about the effectiveness and efficiency of each approach.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Optical Music Recognition is a field of research that automates the reading of musical scores so as to transcribe their content into a structured digital format. When dealing with music manuscripts, the traditional workflow establishes separate stages of detection and classification of musical symbols. In the latter, most of the research has focused on detecting musical glyphs, ignoring that the meaning of a musical symbol is defined by two components: its glyph and its position within the staff. In this paper we study how to perform both glyph and position classification of handwritten musical symbols in early music manuscripts written in white Mensural notation, a common notation system used for the most part of the XVI and XVII centuries. We make use of Convolutional Neural Networks as the classification method, and we tested several alternatives such as using independent models for each component, combining label spaces, or using both multi-input and multi-output models. Our results on early music manuscripts provide insights about the effectiveness and efficiency of each approach.
Bustos, A.
Extraction of medical knowledge from clinical reports and chest x-rays using machine learning techniques PhD Thesis
2019.
BibTeX | Tags:
@phdthesis{k459,
title = {Extraction of medical knowledge from clinical reports and chest x-rays using machine learning techniques},
author = {A. Bustos},
editor = {A. Pertusa},
year = {2019},
date = {2019-07-01},
urldate = {2019-07-01},
organization = {University of Alicante},
keywords = {},
pubstate = {published},
tppubtype = {phdthesis}
}
Gallego, A. J.; Gil, P.; Pertusa, A.; Fisher, R. B.
Semantic Segmentation of SLAR Imagery with Convolutional LSTM Selectional AutoEncoders Journal Article
In: Remote Sensing, vol. 11, no. 12, pp. 1-22, 2019.
@article{k412,
title = {Semantic Segmentation of SLAR Imagery with Convolutional LSTM Selectional AutoEncoders},
author = {A. J. Gallego and P. Gil and A. Pertusa and R. B. Fisher},
year = {2019},
date = {2019-06-01},
journal = {Remote Sensing},
volume = {11},
number = {12},
pages = {1-22},
abstract = {We present a method to detect maritime oil spills from Side-Looking Airborne Radar (SLAR) sensors mounted on aircraft in order to enable a quick response of emergency services when an oil spill occurs. The proposed approach introduces a new type of neural architecture named Convolutional Long Short Term Memory Selectional AutoEncoders (CMSAE) which allows the simultaneous segmentation of multiple classes such as coast, oil spill and ships. Unlike previous works using full SLAR images, in this work only a few scanlines from the beam-scanning of radar are needed to perform the detection. The main objective is to develop a method that performs accurate segmentation using only the current and previous sensor information, in order to return a real-time response during the flight. The proposed architecture uses a series of CMSAE networks to process in parallel each of the objectives defined as different classes. The output of these networks are given to a machine learning classifier to perform the final detection. Results show that the proposed approach can reliably detect oil spills and other maritime objects in SLAR sequences, outperforming the accuracy of previous state-of-the-art methods and with a response time of only 0.76 s.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
We present a method to detect maritime oil spills from Side-Looking Airborne Radar (SLAR) sensors mounted on aircraft in order to enable a quick response of emergency services when an oil spill occurs. The proposed approach introduces a new type of neural architecture named Convolutional Long Short Term Memory Selectional AutoEncoders (CMSAE) which allows the simultaneous segmentation of multiple classes such as coast, oil spill and ships. Unlike previous works using full SLAR images, in this work only a few scanlines from the beam-scanning of radar are needed to perform the detection. The main objective is to develop a method that performs accurate segmentation using only the current and previous sensor information, in order to return a real-time response during the flight. The proposed architecture uses a series of CMSAE networks to process in parallel each of the objectives defined as different classes. The output of these networks are given to a machine learning classifier to perform the final detection. Results show that the proposed approach can reliably detect oil spills and other maritime objects in SLAR sequences, outperforming the accuracy of previous state-of-the-art methods and with a response time of only 0.76 s.
Calvo-Zaragoza, J.; Rico-Juan, J. R.; Gallego, A. J.
Ensemble classification from deep predictions with test data augmentation Journal Article
In: Soft Computing, 2019, ISSN: 1433-7479.
@article{k409,
title = {Ensemble classification from deep predictions with test data augmentation},
author = {J. Calvo-Zaragoza and J. R. Rico-Juan and A. J. Gallego},
issn = {1433-7479},
year = {2019},
date = {2019-04-01},
urldate = {2019-04-01},
journal = {Soft Computing},
abstract = {Data augmentation has become a standard step to improve the predictive power and robustness of convolutional neural networks by means of the synthetic generation of new samples depicting different deformations. This step has been traditionally considered to improve the network at the training stage. In this work, however, we study the use of data augmentation at classification time. That is, the test sample is augmented, following the same procedure considered for training, and the decision is taken with an ensemble prediction over all these samples. We present comprehensive experimentation with several datasets and ensemble decisions, considering a rather generic data augmentation procedure. Our results show that performing this step is able to boost the original classification, even when the room for improvement is limited.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Data augmentation has become a standard step to improve the predictive power and robustness of convolutional neural networks by means of the synthetic generation of new samples depicting different deformations. This step has been traditionally considered to improve the network at the training stage. In this work, however, we study the use of data augmentation at classification time. That is, the test sample is augmented, following the same procedure considered for training, and the decision is taken with an ensemble prediction over all these samples. We present comprehensive experimentation with several datasets and ensemble decisions, considering a rather generic data augmentation procedure. Our results show that performing this step is able to boost the original classification, even when the room for improvement is limited.
Rizo, D.; Marsden, A.
An MEI-based standard encoding for hierarchical music analyses Journal Article
In: International Journal on Digital Libraries, vol. 20, no. 1, pp. 93–105, 2019, ISSN: 1432-5012.
@article{k402,
title = {An MEI-based standard encoding for hierarchical music analyses},
author = {D. Rizo and A. Marsden},
issn = {1432-5012},
year = {2019},
date = {2019-03-01},
journal = {International Journal on Digital Libraries},
volume = {20},
number = {1},
pages = {93--105},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {article}
}
Calvo-Zaragoza, J.; Gallego, A. J.
A selectional auto-encoder approach for document image binarization Journal Article
In: Pattern Recognition, vol. 86, pp. 37-47, 2019, ISSN: 0031-3203.
Abstract | Links | BibTeX | Tags: GRE16-14
@article{k395,
title = {A selectional auto-encoder approach for document image binarization},
author = {J. Calvo-Zaragoza and A. J. Gallego},
url = {https://grfia.dlsi.ua.es/repositori/grfia/pubs/395/1706.10241.pdf},
issn = {0031-3203},
year = {2019},
date = {2019-01-01},
journal = {Pattern Recognition},
volume = {86},
pages = {37-47},
abstract = {Binarization plays a key role in the automatic information retrieval from document images. This process is usually performed in the first stages of document analysis systems, and serves as a basis for subsequent steps. Hence it has to be robust in order to allow the full analysis workflow to be successful. Several methods for document image binarization have been proposed so far, most of which are based on hand-crafted image processing strategies. Recently, Convolutional Neural Networks have shown an amazing performance in many disparate duties related to computer vision. In this paper we discuss the use of convolutional auto-encoders devoted to learning an end-to-end map from an input image to its selectional output, in which activations indicate the likelihood of pixels to be either foreground or background. Once trained, documents can therefore be binarized by parsing them through the model and applying a global threshold. This approach has proven to outperform existing binarization strategies in a number of document types.},
keywords = {GRE16-14},
pubstate = {published},
tppubtype = {article}
}
Binarization plays a key role in the automatic information retrieval from document images. This process is usually performed in the first stages of document analysis systems, and serves as a basis for subsequent steps. Hence it has to be robust in order to allow the full analysis workflow to be successful. Several methods for document image binarization have been proposed so far, most of which are based on hand-crafted image processing strategies. Recently, Convolutional Neural Networks have shown an amazing performance in many disparate duties related to computer vision. In this paper we discuss the use of convolutional auto-encoders devoted to learning an end-to-end map from an input image to its selectional output, in which activations indicate the likelihood of pixels to be either foreground or background. Once trained, documents can therefore be binarized by parsing them through the model and applying a global threshold. This approach has proven to outperform existing binarization strategies in a number of document types.
Gallego, A. J.; Pertusa, A.; Gil, P.; Fisher, R. B.
Detection of bodies in maritime rescue operations using unmanned aerial vehicles with multispectral cameras Journal Article
In: Journal of Field Robotics, vol. 36, no. 4, pp. 782-796, 2019.
BibTeX | Tags:
@article{k401,
title = {Detection of bodies in maritime rescue operations using unmanned aerial vehicles with multispectral cameras},
author = {A. J. Gallego and A. Pertusa and P. Gil and R. B. Fisher},
year = {2019},
date = {2019-01-01},
journal = {Journal of Field Robotics},
volume = {36},
number = {4},
pages = {782-796},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Alashhab, S.; Gallego, A. J.; Lozano, M. Á.
Hand Gesture Detection with Convolutional Neural Networks Proceedings Article
In: Distributed Computing and Artificial Intelligence, 15th International Conference, pp. 45–52, Springer International Publishing, 2019, ISBN: 978-3-319-94649-8.
@inproceedings{k405,
title = {Hand Gesture Detection with Convolutional Neural Networks},
author = {S. Alashhab and A. J. Gallego and M. Á. Lozano},
isbn = {978-3-319-94649-8},
year = {2019},
date = {2019-01-01},
urldate = {2019-01-01},
booktitle = {Distributed Computing and Artificial Intelligence, 15th International Conference},
pages = {45--52},
publisher = {Springer International Publishing},
abstract = {In this paper, we present a method for locating and recognizing hand gestures from images, based on Deep Learning. Our goal is to provide an intuitive and accessible way to interact with Computer Vision-based mobile applications aimed to assist visually impaired people (e.g. pointing a finger at an object in a real scene to zoom in for a close-up of the pointed object). Initially, we have defined different hand gestures that can be assigned to different actions. After that, we have created a database containing images corresponding to these gestures. Lastly, this database has been used to train Neural Networks with different topologies (testing different input sizes, weight initialization, and data augmentation process). In our experiments, we have obtained high accuracies both in localization (96%-100%) and in recognition (99.45%) with Networks that are appropriate to be ported to mobile devices.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
In this paper, we present a method for locating and recognizing hand gestures from images, based on Deep Learning. Our goal is to provide an intuitive and accessible way to interact with Computer Vision-based mobile applications aimed to assist visually impaired people (e.g. pointing a finger at an object in a real scene to zoom in for a close-up of the pointed object). Initially, we have defined different hand gestures that can be assigned to different actions. After that, we have created a database containing images corresponding to these gestures. Lastly, this database has been used to train Neural Networks with different topologies (testing different input sizes, weight initialization, and data augmentation process). In our experiments, we have obtained high accuracies both in localization (96%-100%) and in recognition (99.45%) with Networks that are appropriate to be ported to mobile devices.
Garay-Maestre, U.; Gallego, A. J.; Calvo-Zaragoza, J.
Data Augmentation via Variational Auto-Encoders Proceedings Article
In: Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications, pp. 29–37, 2019, ISBN: 978-3-030-13469-3.
Abstract | BibTeX | Tags: HispaMus
@inproceedings{k406,
title = {Data Augmentation via Variational Auto-Encoders},
author = {U. Garay-Maestre and A. J. Gallego and J. Calvo-Zaragoza},
isbn = {978-3-030-13469-3},
year = {2019},
date = {2019-01-01},
urldate = {2019-01-01},
booktitle = {Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications},
pages = {29--37},
abstract = {Data augmentation is a widely considered technique to improve the performance of Convolutional Neural Networks during training. This step consists in synthetically generate new labeled data by perturbing the samples of the training set, which is expected to provide more robustness to the learning process. The problem is that the augmentation procedure has to be adjusted manually because the perturbations considered must make sense for the task at issue. In this paper we propose the use of Variational Auto-Encoders (VAEs) to generate new synthetic samples, instead of resorting to heuristic strategies. VAEs are powerful generative models that learn a parametric latent space of the input domain from which new samples can be generated. In our experiments over the well-known MNIST dataset, the data augmentation by VAEs improves the base results, yet to a lesser extent of that obtained by a well-adjusted conventional data augmentation. However, the combination of both conventional and VAE-guided data augmentations outperforms all the results, thereby demonstrating the goodness of our proposal.},
keywords = {HispaMus},
pubstate = {published},
tppubtype = {inproceedings}
}
Data augmentation is a widely considered technique to improve the performance of Convolutional Neural Networks during training. This step consists in synthetically generate new labeled data by perturbing the samples of the training set, which is expected to provide more robustness to the learning process. The problem is that the augmentation procedure has to be adjusted manually because the perturbations considered must make sense for the task at issue. In this paper we propose the use of Variational Auto-Encoders (VAEs) to generate new synthetic samples, instead of resorting to heuristic strategies. VAEs are powerful generative models that learn a parametric latent space of the input domain from which new samples can be generated. In our experiments over the well-known MNIST dataset, the data augmentation by VAEs improves the base results, yet to a lesser extent of that obtained by a well-adjusted conventional data augmentation. However, the combination of both conventional and VAE-guided data augmentations outperforms all the results, thereby demonstrating the goodness of our proposal.