一作ECCV 2026SpotlightHumanOmni-Speaker: Identifying Who said What and WhenDetao Bai · Xihan Wei · Zhiheng Ma多模态社会与人本Multimodal, Video & Document UnderstandingOpenReview