Visual Anomaly Detection under Complex View-Illumination Interplay: A Large-Scale Benchmark

TL;DR

This paper introduces M2AD, a large-scale benchmark dataset designed to evaluate the robustness of visual anomaly detection methods under complex view and illumination variations, highlighting current methods' limitations.

Contribution

We present M2AD, a comprehensive benchmark with nearly 120,000 images across multiple view-illumination configurations to assess and improve VAD robustness in real-world scenarios.

Findings

State-of-the-art VAD methods perform poorly on M2AD.

View-illumination interplay significantly affects defect visibility.

Benchmark facilitates development of more robust VAD techniques.

Abstract

The practical deployment of Visual Anomaly Detection (VAD) systems is hindered by their sensitivity to real-world imaging variations, particularly the complex interplay between viewpoint and illumination which drastically alters defect visibility. Current benchmarks largely overlook this critical challenge. We introduce Multi-View Multi-Illumination Anomaly Detection (M2AD), a new large-scale benchmark comprising 119,880 high-resolution images designed explicitly to probe VAD robustness under such interacting conditions. By systematically capturing 999 specimens across 10 categories using 12 synchronized views and 10 illumination settings (120 configurations total), M2AD enables rigorous evaluation. We establish two evaluation protocols: M2AD-Synergy tests the ability to fuse information across diverse configurations, and M2AD-Invariant measures single-image robustness against realistic view-illumination effects. Our extensive benchmarking shows that state-of-the-art VAD methods struggle significantly on M2AD, demonstrating the profound challenge posed by view-illumination interplay. This benchmark serves as an essential tool for developing and validating VAD methods capable of overcoming real-world complexities. Our full dataset and test suite will be released at https://hustcyq.github.io/M2AD to facilitate the field.