AI RESEARCH
Point-Supervised Skeleton-Based Human Action Segmentation
arXiv CS.CV
•
ArXi:2603.06201v1 Announce Type: new Skeleton-based temporal action segmentation is a fundamental yet challenging task, playing a crucial role in enabling intelligent systems to perceive and respond to human activities. While fully-supervised methods achieve satisfactory performance, they require costly frame-level annotations and are sensitive to ambiguous action boundaries. To address these issues, we