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

POSESTITCH-SLT is a novel method introduced for end-to-end sign language translation that employs a pre-training scheme inspired by linguistic templates. This approach is designed to tackle the significant challenge of limited large-scale, aligned datasets in the field of sign language translation. By leveraging linguistic insights, POSESTITCH-SLT aims to improve the translation process, making it more effective despite data constraints. The method has been evaluated across two different sign language datasets, where it demonstrated promising results in translation performance. This suggests that the linguistically inspired pose-stitching technique could enhance the accuracy and reliability of sign language translation systems. Overall, POSESTITCH-SLT represents a meaningful advancement in addressing data scarcity while improving translation quality in sign language applications.