Research map / Video / Video segmentation and tracking
Hierarchical LSTM for Sign Language Translation
Dan Guo, Wengang Zhou, Houqiang Li, Meng Wang
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 |
|---|---|---|---|
| hierarchical variable-length encoder instead of fixed-length stacked LSTM (S2VT) | encoder-decoder SLT · CSL Split I | precision: 0.897 -> 0.924 | Improved |
| add temporal attention-aware weighting (ablation) | HLSTM · CSL Split I | precision: 0.924 -> 0.929 | Improved |
| add temporal attention-aware weighting (ablation) | HLSTM · CSL Split II | Acc-w: 0.482 -> 0.506 | Improved |
| add online key clip mining instead of systematic sampling (ablation) | HLSTM · CSL Split I | precision: 0.910 -> 0.924 | Improved |
| add key clip mining instead of systematic sampling (ablation) | HLSTM · CSL Split II | WER: 0.630 -> 0.662 | Hurt |
| use C3D features instead of VGG (ablation) | HLSTM · CSL Split I | precision: 0.446 -> 0.852 | Improved |
| increase LSTM hidden size 256 to 1000 (ablation) | HLSTM · CSL Split I | precision: 0.852 -> 0.924 | Improved |
| mean pooling instead of max pooling of less-important clips (ablation) | HLSTM · CSL Split I vs II | precision / Acc-w: 0.912 -> 0.924; 0.482 -> 0.458 | Mixed |
| compact encoder length n''=21 instead of 66 (ablation) | S2VT · CSL Split I | precision: 0.850 -> 0.897 | Improved |
Closest papers by meaning
- LLMs are Good Sign Language Translators (CVPR 2024 · similarity 0.89)
- Scaling Sign Language Translation (NeurIPS 2024 · similarity 0.89)
- Lost in Translation, Found in Embeddings: Sign Language Translation and Alignment (NeurIPS 2026 · similarity 0.89)
- Conditional Variational Autoencoder for Sign Language Translation with Cross-Modal Alignment (AAAI 2024 · similarity 0.88)
- Bridging Sign and Spoken Languages: Pseudo Gloss Generation for Sign Language Translation (NeurIPS 2025 · similarity 0.88)
- Lost in Translation, Found in Context: Sign Language Translation with Contextual Cues (CVPR 2025 · similarity 0.88)
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.
