Skip to content
arXiv cs.AI · Papers

VQ-VAD: Vector-quantized Motion Representation Learning for Human-centric Video Anomaly Detection

arXiv:2608.05069v1 Announce Type: cross Abstract: Video Anomaly Detection (VAD) is inherently challenging due to the scarcity of anomalies and the large visual variability in surveillance footage, including changes in lighting, viewpoint, and human appearance. To mitigate visual noise and address privacy concerns, rece