Get Started
Home
Topics
Search
Library
Research questionHow can video-based sign dictionary retrieval generalize across signers and unseen languages?A query video may depict the same sign differently across people, while systems trained for closed-set recognition can fail when the corpus or language changes. The central difficulty is matching the video to the correct dictionary entry without language- or signer-specific adaptation.
AI
Computer Vision
Image & Video Processing
Information Retrieval
Machine Learning
Latest papersRecent research connected to this question, newest first.SignSeek: Learning Transferable Representations for Sign Dictionary RetrievalThe evidence concerns retrieval from query videos, using representations pretrained on 266,000 samples across multiple sign languages. It reports cross-corpus results on ASL-Citizen, WLASL, and NMFs-CSL, zero-shot transfer to unseen British Sign Language, and transfer to isolated sign recognition and subtitle alignment without downstream fine-tuning.research paper · Sep 3, 2026
Related questions
How can sign recognition from continuous signing identify unseen signs without gloss-annotated vocabularies?How can sign language translation models capture asynchronous lip cues and recognize fingerspelled terms without detailed supervision?How can video-language models capture the distribution of human interpretations of dynamic facial expressions?How can we build a public surgical video-language corpus that supports reasoning across open and minimally invasive procedures?