Halil Helvaci

New active learning framework targets video action segmentation boundaries

Researchers have developed a new active learning framework called B-ACT for temporal action segmentation in videos. This method focuses on efficiently annotating crucial boundary regions where action transitions occur, …