Shuting He

Event-Aware Instructed Assistant for Referring Video Segmentation

Jinyu Liu, Henghui Ding, Shuting He, Yu-Gang Jiang

Existing referring video segmentation methods often treat a video as a single event consisting of multiple images, overlooking the fact that a video typically contains multiple distinct events. Under such a mechanism, the model needs to directly understand all the complex content…Read more