Research map / Video / Video segmentation and tracking
Event Stream-based Sign Language Translation: A High-Definition Benchmark Dataset and A Novel Baseline
Shiao Wang, Xiao Wang, Duoqing Yang, Yao Rong, Fuling Wang, Jianing Li, Lin Zhu, Bo Jiang
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 spatial detail branch with Mamba aggregation and GSTF to temporal branch (ablation) | EvSLT · Event-CSL | BLEU-4: 48.20 -> 49.80 | Improved |
| add graph-guided spatiotemporal fusion (GSTF) (ablation) | EvSLT · Event-CSL | BLEU-4: 48.84 -> 49.80 | Improved |
| add Mamba memory aggregation (ablation) | EvSLT · Event-CSL | BLEU-4: 49.20 -> 49.80 | Improved |
| use hypergraph in fusion module (ablation) | GSTF · Event-CSL | BLEU-4: 49.07 -> 49.80 | Improved |
| use GSTF instead of concatenation fusion (ablation) | EvSLT · Event-CSL | BLEU-4: 47.00 -> 49.80 | Improved |
| use GSTF instead of cross-attention fusion (ablation) | EvSLT · Event-CSL | BLEU-4: 48.84 -> 49.80 | Improved |
| use all L input frames instead of L/8 (ablation) | EvSLT · Event-CSL | BLEU-4: 10.62 -> 49.80 | Improved |
| use 224x224 instead of 320x320 input resolution (ablation) | EvSLT · Event-CSL | BLEU-4: 49.68 -> 49.80 | Hurt |
| use 224x224 instead of 112x112 input resolution (ablation) | EvSLT · Event-CSL | BLEU-4: 31.72 -> 49.80 | Improved |
| use two Mamba layers instead of one (ablation) | EvSLT · Event-CSL | accuracy: 1.33% decrease with single layer | Improved |
| EvSLT vs GFSLT on PHOENIX-2014T-Event | EvSLT · PHOENIX-2014T-Event | BLEU-4: 14.72 -> 16.48 | Improved |
| EvSLT vs GFSLT on CSL-Daily-Event | EvSLT · CSL-Daily-Event | ROUGE-L: 20.81 -> 27.06 | Improved |
Closest papers by meaning
- Lost in Translation, Found in Embeddings: Sign Language Translation and Alignment (NeurIPS 2026 · similarity 0.88)
- Scaling Sign Language Translation (NeurIPS 2024 · similarity 0.88)
- KD-MSLRT: Lightweight Sign Language Recognition Model Based on Mediapipe and 3D to 1D Knowledge Distillation (AAAI 2025 · similarity 0.87)
- Uni-Sign: Toward Unified Sign Language Understanding at Scale (ICLR 2025 · similarity 0.87)
- Lost in Translation, Found in Context: Sign Language Translation with Contextual Cues (CVPR 2025 · similarity 0.87)
- 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.
