Overview
Note: ASL interpreters will be available throughout the workshop.
- Advance generative approaches for sign language understanding and generation, including high-quality translations, realistic and expressive digital signer synthesis, and expansion of low-resource sign datasets.
- Leverage modern sequential and diffusion-based models to address realism, controllability, and data scarcity in low-resource sign language settings.
- Promote responsible and inclusive AI systems that respect linguistic structure and meaningfully support deaf and hard-of-hearing communities.
Welcome to GenSign: The 1st Workshop on Generative AI for Sign Language at CVPR 2026!
Recent advances in sequential generative models and diffusion-based approaches offer a promising pathway to produce high-quality translations, synthesize realistic and expressive digital signers, expand low-resource sign datasets, and foster more inclusive communication between deaf and hearing communities.
This workshop brings together researchers from computer vision, natural language processing, linguistics, and accessibility studies to explore the frontiers of generative modeling for sign language and to foster responsible, human-centered AI systems that understand and communicate through this uniquely visual language. Specifically, we aim to: