Research map / Video / Video segmentation and tracking
Hand-Model-Aware Sign Language Recognition
Hezhen Hu, Wengang Zhou, Houqiang Li
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 |
|---|---|---|---|
| Add hand-model predictions to full-RGB predictions through late fusion (ablation) | Sign language recognition · NMFs-CSL | top-1 accuracy: 62.1 -> 75.6 | Improved |
| Add hand-model predictions to full-keypoint predictions through late fusion (ablation) | Sign language recognition · NMFs-CSL | top-1 accuracy: 59.9 -> 71.7 | Improved |
| Use MANO-generated 3D hand joints rather than OpenPose 2D hand joints (ablation) | Hand-based ST-GCN recognition · NMFs-CSL | top-1 accuracy: 59.9 -> 71.7 | Improved |
| Add regularization loss to classification loss (ablation) | Hand-model-aware recognition · NMFs-CSL | top-1 accuracy: 61.5 -> 62.0 | Improved |
| Add spatial consistency loss between projected 3D joints and detected 2D joints (ablation) | Hand-model-aware recognition · NMFs-CSL | top-1 accuracy: 62.0 -> 64.0 | Improved |
| Add temporal consistency loss on predicted 3D poses (ablation) | Hand-model-aware recognition · NMFs-CSL | top-1 accuracy: 64.0 -> 64.7 | Improved |
| Use hand-model predictions without late fusion | Sign language recognition · NMFs-CSL | top-1 accuracy: 64.7; GLE-Net 69.0 | Mixed |
| Fuse hand-model predictions with full-RGB predictions (ablation) | Sign language recognition · SLR500 | accuracy: 95.1 -> 98.3 | Improved |
| Fuse hand-model predictions with full-RGB predictions (ablation) | Sign language recognition · WLASL | per-instance top-1 accuracy: 32.48 -> 51.39 | Improved |
Closest papers by meaning
- SignBERT: Pre-Training of Hand-Model-Aware Representation for Sign Language Recognition (ICCV 2021 · similarity 0.94)
- SignBERT+: Hand-Model-Aware Self-Supervised Pre-Training for Sign Language Understanding (TPAMI 2023 · similarity 0.94)
- Scaling up Multimodal Pre-Training for Sign Language Understanding (TPAMI 2025 · similarity 0.90)
- Natural Language-Assisted Sign Language Recognition (CVPR 2023 · similarity 0.89)
- Uni-Sign: Toward Unified Sign Language Understanding at Scale (ICLR 2025 · similarity 0.86)
- Lost in Translation, Found in Embeddings: Sign Language Translation and Alignment (NeurIPS 2026 · similarity 0.86)
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.
