Research map / Video / Video generation
Pose Guided Fine-Grained Sign Language Video Generation
Tongkai Shi, Lianyu Hu, Fanhua Shang, Jichao Feng, liu peidong, Wei Feng
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 pose features to the TPSMM baseline by direct summation (ablation) | TPSMM · PHOENIX-2014T test | WER (%): 55.2% -> 44.0% | Improved |
| replace direct pose addition with cross-attention Pose Fusion Module (ablation) | PGMM · PHOENIX-2014T test | WER (%): 44.0% -> 40.1% | Improved |
| replace direct pose addition with PFM (ablation) | PGMM · PHOENIX-2014T test | FVD: 140.957 -> 135.244 | Improved |
| add Pose Distance Loss on top of PFM (ablation) | PGMM · PHOENIX-2014T test | TCD: 0.126 -> 0.105 | Improved |
| add Feature Alignment Loss on top of PFM (ablation) | PGMM · PHOENIX-2014T test | L1: 0.01396 -> 0.01375 | Improved |
| add Feature Alignment Loss on top of PFM (ablation) | PGMM · PHOENIX-2014T test | WER (%): 40.1% -> 40.6% | No clear effect |
| full PGMM vs TPSMM on video reconstruction | PGMM · PHOENIX-2014T test | WER (%): 55.2% -> 39.7% | Improved |
| full PGMM vs TPSMM on WLASL-2000 temporal consistency | PGMM · WLASL-2000 | TCD: 0.064 -> 0.067 | Hurt |
| full PGMM vs TPSMM on CSL-Daily | PGMM · CSL-Daily test | FVD: 98.778 -> 82.974 | Improved |
Closest papers by meaning
- LLM Is a Good Conditioner: End-to-End Sign Language Video Generation with VQ-Diffusion (NeurIPS 2026 · similarity 0.89)
- Advanced Sign Language Video Generation with Compressed and Quantized Multi-Condition Tokenization (NeurIPS 2025 · similarity 0.88)
- SignRefine: Adapting Foundational Video Models for Sign Language Generation (ECCV 2026 · similarity 0.88)
- Discrete to Continuous: Generating Smooth Transition Poses from Sign Language Observations (CVPR 2025 · similarity 0.87)
- Sign-IDD: Iconicity Disentangled Diffusion for Sign Language Production (AAAI 2025 · similarity 0.86)
- Lost in Translation, Found in Embeddings: Sign Language Translation and Alignment (NeurIPS 2026 · similarity 0.85)
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.
