Abstract: In Continuous Sign Language Recognition (CSLR), frame-wise features are often affected by occlusion and motion blur, leading to unstable predictions. To address this issue, we propose ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results