Collaborative Discrepancy Optimization for Reliable Image Anomaly Localization

doi:10.1109/TII.2023.3241579

Collaborative Discrepancy Optimization for Reliable Image Anomaly Localization

2023-11-01

25

OA

AI

DOI：10.1109/TII.2023.3241579

Original

Original request for help

Save

Abstract

En 中文

Most unsupervised image anomaly localization methods suffer from overgeneralization because of the high generalization abilities of convolutional neural networks, leading to unreliable predictions. To mitigate the overgeneralization, this article proposes to collaboratively optimize normal and abnormal feature distributions with the assistance of synthetic anomalies, namely collaborative discrepancy optimization (CDO). CDO introduces a margin optimization module and an overlap optimization module to optimize the two key factors determining the localization performance, i.e., the margin and the overlap between the discrepancy distributions (DDs) of normal and abnormal samples. With CDO, a large margin and a small overlap between normal and abnormal DDs are obtained, and the prediction reliability is boosted. Experiments on MVTec2D and MVTec3D show that CDO effectively mitigates the overgeneralization and achieves great anomaly localization performance with real-time computation efficiency. A real-world automotive plastic parts inspection application further demonstrates the capability of the proposed CDO.

Keywords:

Anomaly localization

collaborative discrepancy

optimization (CDO)

computer vision

defect detection.

Collaborative Discrepancy Optimization for Reliable Image Anomaly Localization

Abstract

Journal

Organization