1 paper
Qihang Zhou, Binbin Gao, Guansong Pang +3
Adapting CLIP for anomaly detection on unseen objects has shown strong potential in a zero-shot manner. However, existing methods typically rely on a single textual space to align…