POSESTITCH-SLT: Linguistically Inspired Pose-Stitching for End-to-End Sign Language Translation

TL;DR

This paper introduces POSESTITCH-SLT, a novel pre-training approach inspired by linguistic templates, which significantly improves pose-based sign language translation performance in low-resource settings using synthetic supervision.

Contribution

The paper presents a new pre-training scheme that leverages linguistic-template-based sentence generation to enhance neural sign language translation models.

Findings

BLEU-4 scores improved on How2Sign and iSign datasets

Template-driven synthetic supervision outperforms prior methods

Effective in low-resource sign language translation scenarios

Abstract

Sign language translation remains a challenging task due to the scarcity of large-scale, sentence-aligned datasets. Prior arts have focused on various feature extraction and architectural changes to support neural machine translation for sign languages. We propose POSESTITCH-SLT, a novel pre-training scheme that is inspired by linguistic-templates-based sentence generation technique. With translation comparison on two sign language datasets, How2Sign and iSign, we show that a simple transformer-based encoder-decoder architecture outperforms the prior art when considering template-generated sentence pairs in training. We achieve BLEU-4 score improvements from 1.97 to 4.56 on How2Sign and from 0.55 to 3.43 on iSign, surpassing prior state-of-the-art methods for pose-based gloss-free translation. The results demonstrate the effectiveness of template-driven synthetic supervision in low-resource sign language settings.