ECCV 2022 Sign Spotting Challenge: Dataset, Design and Results

Manuel Vázquez Enríquez; José L.Alba Castro; Laura Docio Fernandez; Julio C.S. Jacques Junior; Sergio Escalera

doi:10.1007/978-3-031-25085-9_13

ECCV 2022 Sign Spotting Challenge: Dataset, Design and Results

Manuel Vázquez Enríquez^*, José L.Alba Castro, Laura Docio Fernandez, Julio C.S. Jacques Junior, Sergio Escalera

^*Kontaktforfatter

Publikation: Bidrag til bog/antologi/rapport/konference proceeding › Konferenceartikel i proceeding › Forskning › peer review

1 Citationer (Scopus)

Abstract

The ECCV 2022 Sign Spotting Challenge focused on the problem of fine-grain sign spotting for continuous sign language recognition. We have released and made publicly available a new dataset of Spanish sign language of around 10 h of video data in the health domain performed by 7 deaf people and 3 interpreters. The added value of this dataset over existing ones is the frame-level precise annotation of 100 signs with their corresponding glosses and variants made by sign language experts. This paper summarizes the design and results of the challenge, which attracted 79 participants, contextualizing the problem and defining the dataset, protocols and baseline models, as well as discussing top-winning solutions and future directions on the topic.

Originalsprog	Engelsk
Titel	Computer Vision – ECCV 2022 Workshops, Proceedings
Redaktører	Leonid Karlinsky, Tomer Michaeli, Ko Nishino
Antal sider	18
Forlag	Springer Science+Business Media
Publikationsdato	2023
Sider	225-242
ISBN (Trykt)	9783031250842
DOI	https://doi.org/10.1007/978-3-031-25085-9_13
Status	Udgivet - 2023
Udgivet eksternt	Ja
Begivenhed	17th European Conference on Computer Vision, ECCV 2022 - Tel Aviv, Israel Varighed: 23 okt. 2022 → 27 okt. 2022

Konference

Konference	17th European Conference on Computer Vision, ECCV 2022
Land/Område	Israel
By	Tel Aviv
Periode	23/10/2022 → 27/10/2022

Navn	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Vol/bind	13808 LNCS
ISSN	0302-9743

Bibliografisk note

Publisher Copyright:
© 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Adgang til dokumentet

10.1007/978-3-031-25085-9_13

AUB Link

Søg efter materialet i Aalborg Universitetsbiblioteks søgemaskine

Andre filer og links

Link to publication in Scopus

Citationsformater

Vázquez Enríquez, Manuel ; Castro, José L.Alba ; Fernandez, Laura Docio et al. / ECCV 2022 Sign Spotting Challenge : Dataset, Design and Results. Computer Vision – ECCV 2022 Workshops, Proceedings. red. / Leonid Karlinsky ; Tomer Michaeli ; Ko Nishino. Springer Science+Business Media, 2023. s. 225-242 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Bind 13808 LNCS).

@inproceedings{591dbc11a5b240a2bf458b8e640eb7c5,

title = "ECCV 2022 Sign Spotting Challenge: Dataset, Design and Results",

abstract = "The ECCV 2022 Sign Spotting Challenge focused on the problem of fine-grain sign spotting for continuous sign language recognition. We have released and made publicly available a new dataset of Spanish sign language of around 10 h of video data in the health domain performed by 7 deaf people and 3 interpreters. The added value of this dataset over existing ones is the frame-level precise annotation of 100 signs with their corresponding glosses and variants made by sign language experts. This paper summarizes the design and results of the challenge, which attracted 79 participants, contextualizing the problem and defining the dataset, protocols and baseline models, as well as discussing top-winning solutions and future directions on the topic.",

keywords = "3DCNN, Sign language, Sign spotting, ST-GCN",

author = "{V{\'a}zquez Enr{\'i}quez}, Manuel and Castro, {Jos{\'e} L.Alba} and Fernandez, {Laura Docio} and {Jacques Junior}, {Julio C.S.} and Sergio Escalera",

note = "Funding Information: Acknowledgments. This work has been supported by the Spanish projects PID2019-105093GB-I00 and RTI2018-101372-B-I00, by ICREA under the ICREA Academia programme and by the Xunta de Galicia and ERDF through the Consolidated Strategic Group AtlanTTic (2019–2022). Manuel V{\'a}zquez is funded by the Spanish Ministry of Science and Innovation through the predoc grant PRE2019-088146. Publisher Copyright: {\textcopyright} 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 17th European Conference on Computer Vision, ECCV 2022 ; Conference date: 23-10-2022 Through 27-10-2022",

year = "2023",

doi = "10.1007/978-3-031-25085-9_13",

language = "English",

isbn = "9783031250842",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science+Business Media",

pages = "225--242",

editor = "Leonid Karlinsky and Tomer Michaeli and Ko Nishino",

booktitle = "Computer Vision – ECCV 2022 Workshops, Proceedings",

address = "United States",

}

Vázquez Enríquez, M, Castro, JLA, Fernandez, LD, Jacques Junior, JCS & Escalera, S 2023, ECCV 2022 Sign Spotting Challenge: Dataset, Design and Results. i L Karlinsky, T Michaeli & K Nishino (red), Computer Vision – ECCV 2022 Workshops, Proceedings. Springer Science+Business Media, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), bind 13808 LNCS, s. 225-242, 17th European Conference on Computer Vision, ECCV 2022, Tel Aviv, Israel, 23/10/2022. https://doi.org/10.1007/978-3-031-25085-9_13

ECCV 2022 Sign Spotting Challenge: Dataset, Design and Results. / Vázquez Enríquez, Manuel; Castro, José L.Alba; Fernandez, Laura Docio et al.
Computer Vision – ECCV 2022 Workshops, Proceedings. red. / Leonid Karlinsky; Tomer Michaeli; Ko Nishino. Springer Science+Business Media, 2023. s. 225-242 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Bind 13808 LNCS).

Publikation: Bidrag til bog/antologi/rapport/konference proceeding › Konferenceartikel i proceeding › Forskning › peer review

TY - GEN

T1 - ECCV 2022 Sign Spotting Challenge

T2 - 17th European Conference on Computer Vision, ECCV 2022

AU - Vázquez Enríquez, Manuel

AU - Castro, José L.Alba

AU - Fernandez, Laura Docio

AU - Jacques Junior, Julio C.S.

AU - Escalera, Sergio

N1 - Funding Information: Acknowledgments. This work has been supported by the Spanish projects PID2019-105093GB-I00 and RTI2018-101372-B-I00, by ICREA under the ICREA Academia programme and by the Xunta de Galicia and ERDF through the Consolidated Strategic Group AtlanTTic (2019–2022). Manuel Vázquez is funded by the Spanish Ministry of Science and Innovation through the predoc grant PRE2019-088146. Publisher Copyright: © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

PY - 2023

Y1 - 2023

N2 - The ECCV 2022 Sign Spotting Challenge focused on the problem of fine-grain sign spotting for continuous sign language recognition. We have released and made publicly available a new dataset of Spanish sign language of around 10 h of video data in the health domain performed by 7 deaf people and 3 interpreters. The added value of this dataset over existing ones is the frame-level precise annotation of 100 signs with their corresponding glosses and variants made by sign language experts. This paper summarizes the design and results of the challenge, which attracted 79 participants, contextualizing the problem and defining the dataset, protocols and baseline models, as well as discussing top-winning solutions and future directions on the topic.

AB - The ECCV 2022 Sign Spotting Challenge focused on the problem of fine-grain sign spotting for continuous sign language recognition. We have released and made publicly available a new dataset of Spanish sign language of around 10 h of video data in the health domain performed by 7 deaf people and 3 interpreters. The added value of this dataset over existing ones is the frame-level precise annotation of 100 signs with their corresponding glosses and variants made by sign language experts. This paper summarizes the design and results of the challenge, which attracted 79 participants, contextualizing the problem and defining the dataset, protocols and baseline models, as well as discussing top-winning solutions and future directions on the topic.

KW - 3DCNN

KW - Sign language

KW - Sign spotting

KW - ST-GCN

UR - http://www.scopus.com/inward/record.url?scp=85151337313&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-25085-9_13

DO - 10.1007/978-3-031-25085-9_13

M3 - Article in proceeding

AN - SCOPUS:85151337313

SN - 9783031250842

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 225

EP - 242

BT - Computer Vision – ECCV 2022 Workshops, Proceedings

A2 - Karlinsky, Leonid

A2 - Michaeli, Tomer

A2 - Nishino, Ko

PB - Springer Science+Business Media

Y2 - 23 October 2022 through 27 October 2022

ER -

Vázquez Enríquez M, Castro JLA, Fernandez LD, Jacques Junior JCS, Escalera S. ECCV 2022 Sign Spotting Challenge: Dataset, Design and Results. I Karlinsky L, Michaeli T, Nishino K, red., Computer Vision – ECCV 2022 Workshops, Proceedings. Springer Science+Business Media. 2023. s. 225-242. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Bind 13808 LNCS). doi: 10.1007/978-3-031-25085-9_13