Two-stream Decoder Feature Normality Estimating Network for Industrial Anomaly Detection

TL;DR

This paper introduces a two-stream decoder network with a feature normality estimator for improved industrial anomaly detection, effectively distinguishing normal and abnormal features to enhance detection accuracy.

Contribution

The paper presents a novel two-stream decoder network combined with a feature normality estimator, explicitly optimizing for anomaly distinguishability in industrial images.

Findings

Outperforms state-of-the-art models on benchmark datasets

Effectively separates normal and abnormal features

Reduces reconstruction of abnormal regions

Abstract

Image reconstruction-based anomaly detection has recently been in the spotlight because of the difficulty of constructing anomaly datasets. These approaches work by learning to model normal features without seeing abnormal samples during training and then discriminating anomalies at test time based on the reconstructive errors. However, these models have limitations in reconstructing the abnormal samples due to their indiscriminate conveyance of features. Moreover, these approaches are not explicitly optimized for distinguishable anomalies. To address these problems, we propose a two-stream decoder network (TSDN), designed to learn both normal and abnormal features. Additionally, we propose a feature normality estimator (FNE) to eliminate abnormal features and prevent high-quality reconstruction of abnormal regions. Evaluation on a standard benchmark demonstrated performance better than state-of-the-art models.