1 paper · 1 filter
Yiming Wang, Yonghao Dang, Huilai Li +2
Existing human-object interaction (HOI) motion captioning methods typically describe what happens while referring to the subject using generic terms such as "a person" or "someone"…