Abstract: Sign language translation (SLT) traditionally requires costly human gloss annotations. Recently, gloss-free approaches, which directly generate text from video, have been studied and ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results