Abstract: Continuous sign language recognition (CSLR) requires fine-grained alignment between visual sequences and gloss annotations under weak supervision, which is challenged by modality ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results