Abstract: Current zero-shot human-object interaction detection methods often ... Besides, extra decoders might increase computation costs. Thus, we propose a triple knowledge cues exploration model ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback