Research map / Large language models / Language, safety and interpretability
Automatic Dense Annotation of Large-Vocabulary Sign Language Videos
Liliane Momeni, Hannah Bull, K R Prajwal, Samuel Albanie, Gül Varol, Andrew Zisserman
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 |
|---|---|---|---|
| expand sign vocabulary from 2K to 8K with 670K annotations | I3D feature extractor · SENT-TEST I3D predictions (subtitle independent) | Recall: 25.5 -> 26.3 | Improved |
| replace prior keyword spotter with Transpotter for mouthings | Mouthing annotations · SENT-TEST MLP predictions 8K (subtitle independent) | Recall: 15.1 -> 20.4 | Improved |
| finetune Transpotter on sign language data by bootstrapping (ablation) | Mouthing annotations · SENT-TEST MLP predictions 8K (subtitle independent) | Recall: 20.4 -> 23.6 | Improved |
| add subtitle alignment to mouthing spotting (ablation) | Mouthing annotations · SENT-TEST MLP predictions 8K (subtitle independent) | Recall: 23.6 -> 27.4 | Improved |
| add subtitle alignment to dictionary spotting (ablation) | Dictionary annotations · SENT-TEST MLP predictions 8K (subtitle independent) | Recall: 24.0 -> 24.2 | No clear effect |
| query synonyms for mouthing spotting (ablation) | Mouthing annotations · SENT-TEST MLP predictions 8K (subtitle independent) | Recall: 27.4 -> 25.3 | Hurt |
| query synonyms for dictionary spotting (ablation) | Dictionary annotations · SENT-TEST MLP predictions 8K (subtitle independent) | Recall: 24.2 -> 26.0 | Improved |
| add attention spottings A (ablation) | M* + D* · SENT-TEST MLP predictions 8K (subtitle independent) | Recall: 29.0 -> 29.6 | Improved |
| add pseudo-label spottings P (ablation) | M* + D* + A · SENT-TEST spottings (subtitle dependent) | Recall: 21.9 -> 25.4 | Improved |
| add pseudo-label spottings P (ablation) | M* + D* + A · SENT-TEST MLP predictions 8K | Recall | No clear effect |
| combine dictionary spottings with improved mouthing (ablation) | M* · SENT-TEST MLP predictions 8K (subtitle independent) | Recall: 27.4 -> 29.0 | Improved |
Closest papers by meaning
- Lost in Translation, Found in Embeddings: Sign Language Translation and Alignment (NeurIPS 2026 · similarity 0.90)
- YouTube-SL-25: A Large-Scale, Open-Domain Multilingual Sign Language Parallel Corpus (ICLR 2025 · similarity 0.90)
- Bridging Sign and Spoken Languages: Pseudo Gloss Generation for Sign Language Translation (NeurIPS 2025 · similarity 0.89)
- Lost in Translation, Found in Context: Sign Language Translation with Contextual Cues (CVPR 2025 · similarity 0.89)
- Scaling Sign Language Translation (NeurIPS 2024 · similarity 0.88)
- Sign2GPT: Leveraging Large Language Models for Gloss-Free Sign Language Translation (ICLR 2024 · similarity 0.87)
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.
