Advanced search
1 file | 1.70 MB Add to list

Towards automatic sign language corpus annotation using deep learning

Mathieu De Coster (UGent) , Mieke Van Herreweghe (UGent) and Joni Dambre (UGent)
Author
Organization
Abstract
Sign classification in sign language corpora is a challenging problem that requires large datasets. Unfortunately, only a small portion of those corpora is labeled. To expedite the annotation process, we propose a gloss suggestion system based on deep learning. We improve upon previous research in three ways. Firstly, we use a proven feature extraction method called OpenPose, rather than learning end-to-end. Secondly, we propose a more suitable and powerful network architecture, based on GRU layers. Finally, we exploit domain and task knowledge to further increase the accuracy. We show that we greatly outperform the previous state of the art on the used dataset. Our method can be used for suggesting a top 5 of annotations given a video fragment that is selected by the corpus annotator. We expect that it will expedite the annotation process to the benefit of sign language translation research.
Keywords
sign language recognition, deep learning, human pose estimation

Downloads

  • (...).pdf
    • full text (Published version)
    • |
    • UGent only
    • |
    • PDF
    • |
    • 1.70 MB

Citation

Please use this url to cite or link to this publication:

MLA
De Coster, Mathieu, et al. “Towards Automatic Sign Language Corpus Annotation Using Deep Learning.” 6th Workshop on Sign Language Translation and Avatar Technology, Proceedings, 2019.
APA
De Coster, M., Van Herreweghe, M., & Dambre, J. (2019). Towards automatic sign language corpus annotation using deep learning. In 6th Workshop on Sign Language Translation and Avatar Technology, Proceedings. Hamburg, Germany.
Chicago author-date
De Coster, Mathieu, Mieke Van Herreweghe, and Joni Dambre. 2019. “Towards Automatic Sign Language Corpus Annotation Using Deep Learning.” In 6th Workshop on Sign Language Translation and Avatar Technology, Proceedings.
Chicago author-date (all authors)
De Coster, Mathieu, Mieke Van Herreweghe, and Joni Dambre. 2019. “Towards Automatic Sign Language Corpus Annotation Using Deep Learning.” In 6th Workshop on Sign Language Translation and Avatar Technology, Proceedings.
Vancouver
1.
De Coster M, Van Herreweghe M, Dambre J. Towards automatic sign language corpus annotation using deep learning. In: 6th Workshop on Sign Language Translation and Avatar Technology, Proceedings. 2019.
IEEE
[1]
M. De Coster, M. Van Herreweghe, and J. Dambre, “Towards automatic sign language corpus annotation using deep learning,” in 6th Workshop on Sign Language Translation and Avatar Technology, Proceedings, Hamburg, Germany, 2019.
@inproceedings{8647050,
  abstract     = {Sign classification in sign language corpora is a challenging problem that requires large datasets. Unfortunately, only a small portion of those corpora is labeled. To expedite the annotation process, we propose a gloss suggestion system based on deep learning. We improve upon previous research in three ways. Firstly, we use a proven feature extraction method called OpenPose, rather than learning end-to-end. Secondly, we propose a more suitable and powerful network architecture, based on GRU layers. Finally, we exploit domain and task knowledge to further increase the accuracy.
We show that we greatly outperform the previous state of the art on the used dataset. Our method can be used for suggesting a top 5 of annotations given a video fragment that is selected by the corpus annotator. We expect that it will expedite the annotation process to the benefit of sign language translation research.},
  author       = {De Coster, Mathieu and Van Herreweghe, Mieke and Dambre, Joni},
  booktitle    = {6th Workshop on Sign Language Translation and Avatar Technology, Proceedings},
  keywords     = {sign language recognition,deep learning,human pose estimation},
  language     = {eng},
  location     = {Hamburg, Germany},
  pages        = {3},
  title        = {Towards automatic sign language corpus annotation using deep learning},
  year         = {2019},
}