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.
Cite
Text
Vázquez-Enríquez et al. "ECCV 2022 Sign Spotting Challenge: Dataset, Design and Results." European Conference on Computer Vision Workshops, 2022. doi:10.1007/978-3-031-25085-9_13Markdown
[Vázquez-Enríquez et al. "ECCV 2022 Sign Spotting Challenge: Dataset, Design and Results." European Conference on Computer Vision Workshops, 2022.](https://mlanthology.org/eccvw/2022/vazquezenriquez2022eccvw-eccv/) doi:10.1007/978-3-031-25085-9_13BibTeX
@inproceedings{vazquezenriquez2022eccvw-eccv,
title = {{ECCV 2022 Sign Spotting Challenge: Dataset, Design and Results}},
author = {Vázquez-Enríquez, Manuel and Alba-Castro, José Luis and Fernández, Laura Docío and Júnior, Júlio C. S. Jacques and Escalera, Sergio},
booktitle = {European Conference on Computer Vision Workshops},
year = {2022},
pages = {225-242},
doi = {10.1007/978-3-031-25085-9_13},
url = {https://mlanthology.org/eccvw/2022/vazquezenriquez2022eccvw-eccv/}
}