FAIR: Frequency-aware Image Restoration for Industrial Visual Anomaly Detection

Tongkun Liu, Bing Li, Xiao Du, Bingke Jiang, Leqi Geng +2 more
9/13/2023
cs.CV

Official code: liutongkun/FAIR (46 stars, last commit 2025-08-23).

Abstract

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.

View on arXivView PDF

Code Implementations(1)

liutongkun/FAIROfficial100%

Frequency-aware Image Restoration for Industrial Visual anomaly detection

462PythonJul 3, 20231 years ago

Similar Papers

Not All Regions Are Equal: Attention-Guided Perturbation Network for Industrial Anomaly Detection

90% match2024

URA-Net: Uncertainty-Integrated Anomaly Perception and Restoration Attention Network for Unsupervised Anomaly Detection

89% match2026

GATE-AD: Graph Attention Network Encoding For Few-Shot Industrial Visual Anomaly Detection

87% match1 repo32026

Produce Once, Utilize Twice for Anomaly Detection

86% match2023

Exploring Intrinsic Normal Prototypes within a Single Image for Universal Anomaly Detection

86% match1 repo3042026

Robust Logical Anomaly Detection under Capture Shifts via Synthetic Anomalies

86% match2026

Cite this paper

@article{liu2023fair,
  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},
  year   = {2023},
  eprint = {2309.07068},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2309.07068v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2309.07068.svg)](https://codeofpaper.com/paper/2309.07068)

Discussion