Patch-wise Auto-Encoder for Visual Anomaly Detection

TL;DR

This paper introduces a patch-wise auto-encoder framework that improves anomaly detection by enhancing reconstruction sensitivity to anomalies, achieving state-of-the-art results on industrial image benchmarks.

Contribution

The novel patch-wise auto-encoder reconstructs images based on spatially distributed features, boosting anomaly detection performance without prior anomaly data.

Findings

Achieves state-of-the-art performance on Mvtec AD benchmark.

Enhances anomaly sensitivity through patch-wise reconstruction.

Proves effectiveness in industrial applications.

Abstract

Anomaly detection without priors of the anomalies is challenging. In the field of unsupervised anomaly detection, traditional auto-encoder (AE) tends to fail based on the assumption that by training only on normal images, the model will not be able to reconstruct abnormal images correctly. On the contrary, we propose a novel patch-wise auto-encoder (Patch AE) framework, which aims at enhancing the reconstruction ability of AE to anomalies instead of weakening it. Each patch of image is reconstructed by corresponding spatially distributed feature vector of the learned feature representation, i.e., patch-wise reconstruction, which ensures anomaly-sensitivity of AE. Our method is simple and efficient. It advances the state-of-the-art performances on Mvtec AD benchmark, which proves the effectiveness of our model. It shows great potential in practical industrial application scenarios.