Fetching the paper…

VCP-CLIP: A visual context prompting model for zero-shot anomaly segmentation · Around