Image reconstruction-based anomaly detection models are widely explored in industrial visual inspection. However, existing models usually suffer from the trade-off between normal reconstruction fidelity and abnormal reconstruction distinguishability, which damages the performance. In this paper, we find that the above trade-off can be better mitigated by leveraging the distinct frequency biases between normal and abnormal reconstruction errors. To this end, we propose Frequency-aware Image Restoration (FAIR), a novel self-supervised image restoration task that restores images from their high-frequency components. It enables precise reconstruction of normal patterns while mitigating unfavorable generalization to anomalies. Using only a simple vanilla UNet, FAIR achieves state-of-the-art performance with higher efficiency on various defect detection datasets. Code: https://github.com/liutongkun/FAIR.
@article{arxiv.2309.07068,
title = {FAIR: Frequency-aware Image Restoration for Industrial Visual Anomaly Detection},
author = {Tongkun Liu and Bing Li and Xiao Du and Bingke Jiang and Leqi Geng and Feiyang Wang and Zhuo Zhao},
journal= {arXiv preprint arXiv:2309.07068},
year = {2025}
}