Papers with offline recognition
Towards Online Continuous Sign Language Recognition and Translation (2024.emnlp-main)
Copied to clipboard
| Challenge: | Existing studies on continuous sign language recognition (CSLR) use offline models with high latency and memory usage. |
| Approach: | They develop a sign dictionary and train an isolated sign language recognition model on the dictionary. |
| Outcome: | The proposed model achieves state-of-the-art on three popular benchmarks across task settings. |