Recognizing Handshapes using Small Datasets

Autores
Cornejo Fandos, Ulises Jeremias; Ríos, Gastón Gustavo; Ronchetti, Franco; Quiroga, Facundo; Hasperué, Waldo; Lanzarini, Laura Cristina
Año de publicación
2019
Idioma
inglés
Tipo de recurso
documento de conferencia
Estado
versión publicada
Descripción
Advances in convolutional neural networks have made possible significant improvements in the state-of-the-art in image classification. However, their success on a particular field rests on the possibility of obtaining labeled data to train networks. Handshape recognition from images, an important subtask of both gesture and sign language recognition, suffers from such a lack of data. Furthermore, hands are highly deformable objects and therefore handshape classification models require larger datasets. We analyze both state of the art models for image classification, as well as data augmentation schemes and specific models to tackle problems with small datasets. In particular, we perform experiments with Wide- DenseNet, a state of the art convolutional architecture and Prototypical Networks, a state of the art few-shot learning meta model. In both cases, we also quantify the impact of data augmentation on accuracy. Our results show that on small and simple data sets such as CIARP, all models and variations of achieve perfect accuracy, and therefore the utility of the data is highly doubtful, despite its having 6000 samples. On the other hand, in small but complex datasets such as LSA16 (800 samples), specialized methods such as Prototypical Networks do have an advantage over other methods. On RWTH, another complex and small dataset with close to 4000 samples, a traditional and state-of-the-art method such as Wide-DenseNet surpasses all other models. Also, data augmentation consistently increases accuracy for Wide-DenseNet, but not fo Prototypical Networks.
XX Workshop de Agentes y Sistemas inteligentes.
Red de Universidades con Carreras en Informática
Materia
Ciencias Informáticas
Sign Language
Hand Shape Recognition
Convolutional Neural Networks
Densenet
Prototypical Networks
Small Datasets
Nivel de accesibilidad
acceso abierto
Condiciones de uso
http://creativecommons.org/licenses/by-nc-sa/4.0/
Repositorio
SEDICI (UNLP)
Institución
Universidad Nacional de La Plata
OAI Identificador
oai:sedici.unlp.edu.ar:10915/90457

id SEDICI_0c9515bea8236a04e3e8225d0da00ca9
oai_identifier_str oai:sedici.unlp.edu.ar:10915/90457
network_acronym_str SEDICI
repository_id_str 1329
network_name_str SEDICI (UNLP)
spelling Recognizing Handshapes using Small DatasetsCornejo Fandos, Ulises JeremiasRíos, Gastón GustavoRonchetti, FrancoQuiroga, FacundoHasperué, WaldoLanzarini, Laura CristinaCiencias InformáticasSign LanguageHand Shape RecognitionConvolutional Neural NetworksDensenetPrototypical NetworksSmall DatasetsAdvances in convolutional neural networks have made possible significant improvements in the state-of-the-art in image classification. However, their success on a particular field rests on the possibility of obtaining labeled data to train networks. Handshape recognition from images, an important subtask of both gesture and sign language recognition, suffers from such a lack of data. Furthermore, hands are highly deformable objects and therefore handshape classification models require larger datasets. We analyze both state of the art models for image classification, as well as data augmentation schemes and specific models to tackle problems with small datasets. In particular, we perform experiments with Wide- DenseNet, a state of the art convolutional architecture and Prototypical Networks, a state of the art few-shot learning meta model. In both cases, we also quantify the impact of data augmentation on accuracy. Our results show that on small and simple data sets such as CIARP, all models and variations of achieve perfect accuracy, and therefore the utility of the data is highly doubtful, despite its having 6000 samples. On the other hand, in small but complex datasets such as LSA16 (800 samples), specialized methods such as Prototypical Networks do have an advantage over other methods. On RWTH, another complex and small dataset with close to 4000 samples, a traditional and state-of-the-art method such as Wide-DenseNet surpasses all other models. Also, data augmentation consistently increases accuracy for Wide-DenseNet, but not fo Prototypical Networks.XX Workshop de Agentes y Sistemas inteligentes.Red de Universidades con Carreras en Informática2019-10info:eu-repo/semantics/conferenceObjectinfo:eu-repo/semantics/publishedVersionObjeto de conferenciahttp://purl.org/coar/resource_type/c_5794info:ar-repo/semantics/documentoDeConferenciaapplication/pdf105-114http://sedici.unlp.edu.ar/handle/10915/90457enginfo:eu-repo/semantics/altIdentifier/isbn/978-987-688-377-1info:eu-repo/semantics/reference/hdl/10915/90359info:eu-repo/semantics/openAccesshttp://creativecommons.org/licenses/by-nc-sa/4.0/Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0)reponame:SEDICI (UNLP)instname:Universidad Nacional de La Platainstacron:UNLP2025-09-29T11:18:37Zoai:sedici.unlp.edu.ar:10915/90457Institucionalhttp://sedici.unlp.edu.ar/Universidad públicaNo correspondehttp://sedici.unlp.edu.ar/oai/snrdalira@sedici.unlp.edu.arArgentinaNo correspondeNo correspondeNo correspondeopendoar:13292025-09-29 11:18:37.81SEDICI (UNLP) - Universidad Nacional de La Platafalse
dc.title.none.fl_str_mv Recognizing Handshapes using Small Datasets
title Recognizing Handshapes using Small Datasets
spellingShingle Recognizing Handshapes using Small Datasets
Cornejo Fandos, Ulises Jeremias
Ciencias Informáticas
Sign Language
Hand Shape Recognition
Convolutional Neural Networks
Densenet
Prototypical Networks
Small Datasets
title_short Recognizing Handshapes using Small Datasets
title_full Recognizing Handshapes using Small Datasets
title_fullStr Recognizing Handshapes using Small Datasets
title_full_unstemmed Recognizing Handshapes using Small Datasets
title_sort Recognizing Handshapes using Small Datasets
dc.creator.none.fl_str_mv Cornejo Fandos, Ulises Jeremias
Ríos, Gastón Gustavo
Ronchetti, Franco
Quiroga, Facundo
Hasperué, Waldo
Lanzarini, Laura Cristina
author Cornejo Fandos, Ulises Jeremias
author_facet Cornejo Fandos, Ulises Jeremias
Ríos, Gastón Gustavo
Ronchetti, Franco
Quiroga, Facundo
Hasperué, Waldo
Lanzarini, Laura Cristina
author_role author
author2 Ríos, Gastón Gustavo
Ronchetti, Franco
Quiroga, Facundo
Hasperué, Waldo
Lanzarini, Laura Cristina
author2_role author
author
author
author
author
dc.subject.none.fl_str_mv Ciencias Informáticas
Sign Language
Hand Shape Recognition
Convolutional Neural Networks
Densenet
Prototypical Networks
Small Datasets
topic Ciencias Informáticas
Sign Language
Hand Shape Recognition
Convolutional Neural Networks
Densenet
Prototypical Networks
Small Datasets
dc.description.none.fl_txt_mv Advances in convolutional neural networks have made possible significant improvements in the state-of-the-art in image classification. However, their success on a particular field rests on the possibility of obtaining labeled data to train networks. Handshape recognition from images, an important subtask of both gesture and sign language recognition, suffers from such a lack of data. Furthermore, hands are highly deformable objects and therefore handshape classification models require larger datasets. We analyze both state of the art models for image classification, as well as data augmentation schemes and specific models to tackle problems with small datasets. In particular, we perform experiments with Wide- DenseNet, a state of the art convolutional architecture and Prototypical Networks, a state of the art few-shot learning meta model. In both cases, we also quantify the impact of data augmentation on accuracy. Our results show that on small and simple data sets such as CIARP, all models and variations of achieve perfect accuracy, and therefore the utility of the data is highly doubtful, despite its having 6000 samples. On the other hand, in small but complex datasets such as LSA16 (800 samples), specialized methods such as Prototypical Networks do have an advantage over other methods. On RWTH, another complex and small dataset with close to 4000 samples, a traditional and state-of-the-art method such as Wide-DenseNet surpasses all other models. Also, data augmentation consistently increases accuracy for Wide-DenseNet, but not fo Prototypical Networks.
XX Workshop de Agentes y Sistemas inteligentes.
Red de Universidades con Carreras en Informática
description Advances in convolutional neural networks have made possible significant improvements in the state-of-the-art in image classification. However, their success on a particular field rests on the possibility of obtaining labeled data to train networks. Handshape recognition from images, an important subtask of both gesture and sign language recognition, suffers from such a lack of data. Furthermore, hands are highly deformable objects and therefore handshape classification models require larger datasets. We analyze both state of the art models for image classification, as well as data augmentation schemes and specific models to tackle problems with small datasets. In particular, we perform experiments with Wide- DenseNet, a state of the art convolutional architecture and Prototypical Networks, a state of the art few-shot learning meta model. In both cases, we also quantify the impact of data augmentation on accuracy. Our results show that on small and simple data sets such as CIARP, all models and variations of achieve perfect accuracy, and therefore the utility of the data is highly doubtful, despite its having 6000 samples. On the other hand, in small but complex datasets such as LSA16 (800 samples), specialized methods such as Prototypical Networks do have an advantage over other methods. On RWTH, another complex and small dataset with close to 4000 samples, a traditional and state-of-the-art method such as Wide-DenseNet surpasses all other models. Also, data augmentation consistently increases accuracy for Wide-DenseNet, but not fo Prototypical Networks.
publishDate 2019
dc.date.none.fl_str_mv 2019-10
dc.type.none.fl_str_mv info:eu-repo/semantics/conferenceObject
info:eu-repo/semantics/publishedVersion
Objeto de conferencia
http://purl.org/coar/resource_type/c_5794
info:ar-repo/semantics/documentoDeConferencia
format conferenceObject
status_str publishedVersion
dc.identifier.none.fl_str_mv http://sedici.unlp.edu.ar/handle/10915/90457
url http://sedici.unlp.edu.ar/handle/10915/90457
dc.language.none.fl_str_mv eng
language eng
dc.relation.none.fl_str_mv info:eu-repo/semantics/altIdentifier/isbn/978-987-688-377-1
info:eu-repo/semantics/reference/hdl/10915/90359
dc.rights.none.fl_str_mv info:eu-repo/semantics/openAccess
http://creativecommons.org/licenses/by-nc-sa/4.0/
Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0)
eu_rights_str_mv openAccess
rights_invalid_str_mv http://creativecommons.org/licenses/by-nc-sa/4.0/
Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0)
dc.format.none.fl_str_mv application/pdf
105-114
dc.source.none.fl_str_mv reponame:SEDICI (UNLP)
instname:Universidad Nacional de La Plata
instacron:UNLP
reponame_str SEDICI (UNLP)
collection SEDICI (UNLP)
instname_str Universidad Nacional de La Plata
instacron_str UNLP
institution UNLP
repository.name.fl_str_mv SEDICI (UNLP) - Universidad Nacional de La Plata
repository.mail.fl_str_mv alira@sedici.unlp.edu.ar
_version_ 1844616059763556352
score 13.070432