Adaptive Deviation Learning for Visual Anomaly Detection with Data Contamination

TL;DR

This paper presents an adaptive deviation learning method for visual anomaly detection that effectively handles data contamination, improving robustness and accuracy in real-world noisy scenarios.

Contribution

The authors introduce a novel deviation learning framework with instance weighting to enhance anomaly detection under contaminated data conditions.

Findings

Outperforms existing methods on MVTec and VisA datasets.

Demonstrates robustness against data contamination.

Achieves higher detection accuracy and stability.

Abstract

Visual anomaly detection targets to detect images that notably differ from normal pattern, and it has found extensive application in identifying defective parts within the manufacturing industry. These anomaly detection paradigms predominantly focus on training detection models using only clean, unlabeled normal samples, assuming an absence of contamination; a condition often unmet in real-world scenarios. The performance of these methods significantly depends on the quality of the data and usually decreases when exposed to noise. We introduce a systematic adaptive method that employs deviation learning to compute anomaly scores end-to-end while addressing data contamination by assigning relative importance to the weights of individual instances. In this approach, the anomaly scores for normal instances are designed to approximate scalar scores obtained from the known prior distribution. Meanwhile, anomaly scores for anomaly examples are adjusted to exhibit statistically significant deviations from these reference scores. Our approach incorporates a constrained optimization problem within the deviation learning framework to update instance weights, resolving this problem for each mini-batch. Comprehensive experiments on the MVTec and VisA benchmark datasets indicate that our proposed method surpasses competing techniques and exhibits both stability and robustness in the presence of data contamination.