LSFB-CONT and LSFB-ISOL: Two New Datasets for Vision-Based Sign Language Recognition

Jerome Fink; Benoît Frénay; Laurence Meurant; Anthony Cleve

LSFB-CONT and LSFB-ISOL: Two New Datasets for Vision-Based Sign Language Recognition

Jerome Fink, Benoît Frénay, Laurence Meurant, Anthony Cleve

Résultats de recherche: Contribution dans un livre/un catalogue/un rapport/dans les actes d'une conférence › Article dans les actes d'une conférence/un colloque

161 Téléchargements (Pure)

Résumé

While significant progress have been made in thefield of Natural Language Processing (NLP), leading the com-mercially available products, Sign Language Recognition (SLR)is still in its infancy. The lack of large-scale sign languagedatasets makes it hard to leverage new Deep Learning methods.In this paper, we introduce LSFB-CONT, a large scale datasetsuited for continuous SLR along with LSFB-ISOL, a subset ofLSFB-CONT for isolated SLR. Baseline SLR experiments areconducted on LSFB-ISOL and the reached accuracy measuresare compared with those obtained on previous datasets. Theresults suggest that state-of-the-art models for action recognitionstill lack sufficient internal representation power to capture thehigh level of variations of a sign language.

langue originale	Anglais
titre	Proceedings of the 2021 International Joint Conference on Neural Networks (IJCNN 2021)
Editeur	IEEE Computer Society Press
Etat de la publication	Publié - 2021

Accès au document

2021_FinkJEtAl_articlemanuscrit soumis, 4,41 MB

Contient cette citation

@inproceedings{a8138b18134042ceb8cfbe3c80a16546,

title = "LSFB-CONT and LSFB-ISOL: Two New Datasets for Vision-Based Sign Language Recognition",

abstract = "While significant progress have been made in thefield of Natural Language Processing (NLP), leading the com-mercially available products, Sign Language Recognition (SLR)is still in its infancy. The lack of large-scale sign languagedatasets makes it hard to leverage new Deep Learning methods.In this paper, we introduce LSFB-CONT, a large scale datasetsuited for continuous SLR along with LSFB-ISOL, a subset ofLSFB-CONT for isolated SLR. Baseline SLR experiments areconducted on LSFB-ISOL and the reached accuracy measuresare compared with those obtained on previous datasets. Theresults suggest that state-of-the-art models for action recognitionstill lack sufficient internal representation power to capture thehigh level of variations of a sign language.",

author = "Jerome Fink and Beno{\^i}t Fr{\'e}nay and Laurence Meurant and Anthony Cleve",

year = "2021",

language = "English",

booktitle = "Proceedings of the 2021 International Joint Conference on Neural Networks (IJCNN 2021)",

publisher = "IEEE Computer Society Press",

}

LSFB-CONT and LSFB-ISOL: Two New Datasets for Vision-Based Sign Language Recognition. / Fink, Jerome ; Frénay, Benoît ; Meurant, Laurence et al.
Proceedings of the 2021 International Joint Conference on Neural Networks (IJCNN 2021). IEEE Computer Society Press, 2021.

Résultats de recherche: Contribution dans un livre/un catalogue/un rapport/dans les actes d'une conférence › Article dans les actes d'une conférence/un colloque

TY - GEN

T1 - LSFB-CONT and LSFB-ISOL

T2 - Two New Datasets for Vision-Based Sign Language Recognition

AU - Fink, Jerome

AU - Frénay, Benoît

AU - Meurant, Laurence

AU - Cleve, Anthony

PY - 2021

Y1 - 2021

N2 - While significant progress have been made in thefield of Natural Language Processing (NLP), leading the com-mercially available products, Sign Language Recognition (SLR)is still in its infancy. The lack of large-scale sign languagedatasets makes it hard to leverage new Deep Learning methods.In this paper, we introduce LSFB-CONT, a large scale datasetsuited for continuous SLR along with LSFB-ISOL, a subset ofLSFB-CONT for isolated SLR. Baseline SLR experiments areconducted on LSFB-ISOL and the reached accuracy measuresare compared with those obtained on previous datasets. Theresults suggest that state-of-the-art models for action recognitionstill lack sufficient internal representation power to capture thehigh level of variations of a sign language.

AB - While significant progress have been made in thefield of Natural Language Processing (NLP), leading the com-mercially available products, Sign Language Recognition (SLR)is still in its infancy. The lack of large-scale sign languagedatasets makes it hard to leverage new Deep Learning methods.In this paper, we introduce LSFB-CONT, a large scale datasetsuited for continuous SLR along with LSFB-ISOL, a subset ofLSFB-CONT for isolated SLR. Baseline SLR experiments areconducted on LSFB-ISOL and the reached accuracy measuresare compared with those obtained on previous datasets. Theresults suggest that state-of-the-art models for action recognitionstill lack sufficient internal representation power to capture thehigh level of variations of a sign language.

M3 - Conference contribution

BT - Proceedings of the 2021 International Joint Conference on Neural Networks (IJCNN 2021)

PB - IEEE Computer Society Press

ER -

LSFB-CONT and LSFB-ISOL: Two New Datasets for Vision-Based Sign Language Recognition

Résumé

Accès au document

Empreinte digitale

Contient cette citation