Research map / Label-efficient and robust learning / Self-supervised and contrastive
Learning Effective Sign Features without Text for Gloss-free Sign Language Translation
Shiwei Gan, Xiao Liu, Yafeng Yin, Nan Liu, Kuizhuang Liu, Desibieer Tuerdaken, Zhiwei Jiang, Lei Xie, Sanglu Lu, Hongkai Wen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use sign-aware DINO pretraining (masked local views for student) instead of standard DINO (ablation) | SignDINO · Phoenix14T test | ROUGE: 39.26 -> 53.79 | Improved |
| pretrain on sign videos with SSL instead of using general-domain DINO frozen (ablation) | GFSLT backbone · Phoenix14T test | ROUGE: 21.32 -> 39.26 | Improved |
| add hand-region local views to global view for student (ablation) | SignDINO · Phoenix14T test | ROUGE: 39.26 -> 49.63 | Improved |
| add face-region view on top of global+hands (ablation) | SignDINO · Phoenix14T test | ROUGE: 49.63 -> 53.79 | Improved |
| swap backbone among DINOv2, DINOv3, Swin, PoolFormer (ablation) | SignDINO · Phoenix14T test | ROUGE: 52.16 to 53.79 | No clear effect |
| initialize from LVD-1689M pretrained weights instead of scratch (ablation) | SignDINO (DINO3-s) · Phoenix14T test | ROUGE: 51.46 -> 52.76 | Improved |
| pretrain on CSL-Daily instead of in-domain Phoenix14T (ablation) | SignDINO · Phoenix14T test | ROUGE: 53.79 -> 41.83 | Hurt |
| increase proportion of pretraining data (ablation) | SignDINO · Phoenix14T | BLEU-4 | Improved |
Closest papers by meaning
- Improving Gloss-free Sign Language Translation by Reducing Representation Density (NeurIPS 2024 · similarity 0.94)
- Bridging Sign and Spoken Languages: Pseudo Gloss Generation for Sign Language Translation (NeurIPS 2025 · similarity 0.93)
- Sign2GPT: Leveraging Large Language Models for Gloss-Free Sign Language Translation (ICLR 2024 · similarity 0.93)
- Scaling Sign Language Translation (NeurIPS 2024 · similarity 0.92)
- Lost in Translation, Found in Embeddings: Sign Language Translation and Alignment (NeurIPS 2026 · similarity 0.91)
- Uni-Sign: Toward Unified Sign Language Understanding at Scale (ICLR 2025 · similarity 0.90)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
