Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Towards Zero-Shot Sign Language Recognition
Yunus Can Bilge, Ramazan Gökberk Cinbiş, Nazlı İkizler-Cinbiş
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-region stream to body stream (ablation) | LLE · ASL-Text ZSL | top-1 accuracy: 11.4 -> 18.0 | Improved |
| add hand-region stream to body stream (ablation) | LLE · MS-ZSSLR-C ZSL | top-1 accuracy: 5.54 -> 5.74 | Improved |
| replace average pooling with bi-LSTM temporal model (ablation) | LLE · ASL-Text ZSL | top-1 accuracy: 18.0 -> 20.9 | Improved |
| replace average pooling with TSM temporal model (ablation) | LLE · ASL-Text and MS-ZSSLR-C ZSL | top-1 accuracy: ASL-Text 18.0 -> 16.8; MS-ZSSLR-C 5.74 -> 7.01 | Mixed |
| replace average pooling with LSTM (ablation) | LLE · ASL-Text ZSL | top-1 accuracy: 18.0 -> 18.2 | No clear effect |
| add attribute embeddings to text embeddings (ablation) | LLE with bi-LSTM · ASL-Text ZSL | top-1 accuracy: 20.9 -> 31.3 | Improved |
| add attribute embeddings to text embeddings (ablation) | LLE with TSM · MS-ZSSLR-C ZSL | top-1 accuracy: 7.01 -> 14.7 | Improved |
| add attributes to text embeddings (top-5) (ablation) | LLE with bi-LSTM · MS-ZSSLR-C ZSL | top-5 accuracy: 19.9 -> 45.2 | Improved |
| use attributes instead of text embeddings (ablation) | LLE with bi-LSTM · ASL-Text ZSL | top-1 accuracy: 20.9 -> 23.7 | Improved |
| add linear reduction of text embedding to d_t=64 (ablation) | LLE with bi-LSTM · ASL-Text ZSL validation | top-1 accuracy: 26.3 to 35.5 | Improved |
| train on wild samples instead of clean samples (ablation) | LLE with TSM · MS-ZSSLR-C clean evaluation | top-1 accuracy: 14.7 -> 10.3 | Hurt |
| use LLE instead of SAE (ablation) | ZSL formulation · ASL-Text ZSL | top-1 accuracy: 8.0 -> 11.4 | Improved |
Closest papers by meaning
- SimZSL: Zero-Shot Learning Beyond a Pre-defined Semantic Embedding Space (IJCV 2025 · similarity 0.87)
- Learning Class Prototypes via Structure Alignment for Zero-Shot Recognition (ECCV 2018 · similarity 0.87)
- An Empirical Study and Analysis of Generalized Zero-Shot Learning for Object Recognition in the Wild (ECCV 2016 · similarity 0.86)
- Domain-Invariant Projection Learning for Zero-Shot Recognition (NeurIPS 2018 · similarity 0.86)
- Zero-Shot Recognition via Direct Classifier Learning with Transferred Samples and Pseudo Labels (AAAI 2017 · similarity 0.86)
- Improved Zero-Shot Classification by Adapting VLMs with Text Descriptions (CVPR 2024 · 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.
