Find location of image anomalies using neural net trained on simulated anomalies
Find location of image anomalies using neural net trained on simulated anomalies
DRAEM — A discriminatively trained reconstruction embedding for surface anomaly detection
arXiv paper abstract https://arxiv.org/abs/2108.07610v1
arXiv PDF paper https://arxiv.org/pdf/2108.07610v1.pdf
Visual surface anomaly detection aims to detect local image regions that significantly deviate from normal appearance.
Recent surface anomaly detection methods rely on generative models … trained only on anomaly-free images, and often require hand-crafted post-processing steps to localize the anomalies
… propose … (DRAEM) … learns a joint representation of an anomalous image and its anomaly-free reconstruction, while simultaneously learning a decision boundary between normal and anomalous examples.
… enables direct anomaly localization without the need for additional complicated post-processing of the network output and can be trained using simple and general anomaly simulations.
… DRAEM outperforms the current state-of-the-art unsupervised methods by a large margin and even delivers detection performance close to the fully-supervised methods on the widely used DAGM surface-defect detection dataset, while substantially outperforming them in localization accuracy.
Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact
Web site with my other posts by category https://morrislee1234.wixsite.com/website
LinkedIn https://www.linkedin.com/in/morris-lee-47877b7b
