Research and industrial anomaly detection
GRD-Net
Generative-reconstructive-discriminative anomaly detection with region-of-interest attention.
Peer-reviewedPublic code
Overview
GRD-Net is one of my main research projects in industrial visual anomaly detection. The method combines generative reconstruction and discriminative segmentation, with region-of-interest attention used to focus the model on the parts of the image that matter for inspection.