SHuBERT: Self-Supervised Sign Language Representation Learning via Multi-Stream Cluster Prediction

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Gueuwou, Shester, Du, Xiaodan, Shakhnarovich, Greg, Livescu, Karen, Liu, Alexander H.
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912461419970560
author Gueuwou, Shester
Du, Xiaodan
Shakhnarovich, Greg
Livescu, Karen
Liu, Alexander H.
author_facet Gueuwou, Shester
Du, Xiaodan
Shakhnarovich, Greg
Livescu, Karen
Liu, Alexander H.
contents Sign language processing has traditionally relied on task-specific models, limiting the potential for transfer learning across tasks. Pre-training methods for sign language have typically focused on either supervised pre-training, which cannot take advantage of unlabeled data, or context-independent (frame or video segment) representations, which ignore the effects of relationships across time in sign language. We introduce SHuBERT (Sign Hidden-Unit BERT), a self-supervised contextual representation model learned from approximately 1,000 hours of American Sign Language video. SHuBERT adapts masked token prediction objectives to multi-stream visual sign language input, learning to predict multiple targets corresponding to clustered hand, face, and body pose streams. SHuBERT achieves state-of-the-art performance across multiple tasks including sign language translation, isolated sign language recognition, and fingerspelling detection.
format Preprint
id arxiv_https___arxiv_org_abs_2411_16765
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle SHuBERT: Self-Supervised Sign Language Representation Learning via Multi-Stream Cluster Prediction
Gueuwou, Shester
Du, Xiaodan
Shakhnarovich, Greg
Livescu, Karen
Liu, Alexander H.
Computation and Language
Computer Vision and Pattern Recognition
Sign language processing has traditionally relied on task-specific models, limiting the potential for transfer learning across tasks. Pre-training methods for sign language have typically focused on either supervised pre-training, which cannot take advantage of unlabeled data, or context-independent (frame or video segment) representations, which ignore the effects of relationships across time in sign language. We introduce SHuBERT (Sign Hidden-Unit BERT), a self-supervised contextual representation model learned from approximately 1,000 hours of American Sign Language video. SHuBERT adapts masked token prediction objectives to multi-stream visual sign language input, learning to predict multiple targets corresponding to clustered hand, face, and body pose streams. SHuBERT achieves state-of-the-art performance across multiple tasks including sign language translation, isolated sign language recognition, and fingerspelling detection.
title SHuBERT: Self-Supervised Sign Language Representation Learning via Multi-Stream Cluster Prediction
topic Computation and Language
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2411.16765