mlsanityData to 13 Oct 2026

A Feature Shuffling and Restoration Strategy for Universal Unsupervised Anomaly Detection

arXiv 2603.22861 ↗First posted 24 Mar 2026cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Unsupervised anomaly detection is vital in industrial fields, with reconstruction-based methods favored for their simplicity and effectiveness. However, reconstruction methods often encounter an identical shortcut issue, where both normal and anomalous regions can be well reconstructed and fail to identify outliers. The severity of this problem increases with the complexity of the normal data distribution. Consequently, existing methods may exhibit excellent detection performance in a specific scenario, but their performance sharply declines when transferred to another scenario. This paper focuses on establishing a universal model applicable to anomaly detection tasks across different settings, termed as universal anomaly detection. In this work, we introduce a novel, straightforward yet efficient framework for universal anomaly detection: ulineFeature ulineShuffling and ulineRestoration (FSR), which can alleviate the identical shortcut issue across different settings. First and foremost, FSR employs multi-scale features with rich semantic information as reconstruction targets, rather than raw image pixels. Subsequently, these multi-scale features are partitioned into non-overlapping feature blocks, which are randomly shuffled and then restored to their original state using a restoration network. This simple paradigm encourages the model to focus more on global contextual information. Additionally, we introduce a novel concept, the shuffling rate, to regulate the complexity of the FSR task, thereby alleviating the identical shortcut across different settings. Furthermore, we provide theoretical explanations for the effectiveness of FSR framework from two perspectives: network structure and mutual information. Extensive experimental results validate the superiority and efficiency of the FSR framework across different settings.Code is available at https://github.com/luow23/FSR.

Key facts. No citing papers in the map yet (data through 13 Oct 2026). Builds on 34 of its 53 references that are in the map. First posted on arXiv 24 Mar 2026.

Builds on

34 of 53 references.

References · 34

  1. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,476
  2. 2016Wide Residual Networkscited by 3,063
  3. 2023Vision Transformers Need Registerscited by 850
  4. 2021Towards Total Recall in Industrial Anomaly Detectioncited by 774
  5. 2020PaDiM: a Patch Distribution Modeling Framework for Anomaly Detection and Localizationcited by 593
  6. 2019Memorizing Normality to Detect Anomaly: Memory-augmented Deep Autoencoder for Unsupervised Anomaly Detectioncited by 562
  7. 2021CutPaste: Self-Supervised Learning for Anomaly Detection and Localizationcited by 530
  8. 2021DRAEM -- A discriminatively trained reconstruction embedding for surface anomaly detectioncited by 511
  9. 2022Anomaly Detection via Reverse Distillation from One-Class Embeddingcited by 420
  10. 2018GANomaly: Semi-Supervised Anomaly Detection via Adversarial Trainingcited by 409
  11. 2019Uninformed Students: Student-Teacher Anomaly Detection with Discriminative Latent Embeddingscited by 386
  12. 2020Multiresolution Knowledge Distillation for Anomaly Detectioncited by 285
22 more
  1. 2022A Unified Model for Multi-class Anomaly Detectioncited by 268
  2. 2018Improving Unsupervised Defect Segmentation by Applying Structural Similarity to Autoencoderscited by 264
  3. 2023SimpleNet: A Simple Network for Image Anomaly Detection and Localizationcited by 256
  4. 2021VT-ADL: A Vision Transformer Network for Image Anomaly Detection and Localizationcited by 222
  5. 2020Patch SVDD: Patch-level SVDD for Anomaly Detection and Segmentationcited by 212
  6. 2020Same Same But DifferNet: Semi-Supervised Defect Detection with Normalizing Flowscited by 212
  7. 2022Registration based Few-Shot Anomaly Detectioncited by 133
  8. 2020Modeling the Distribution of Normal Data in Pre-Trained Deep Features for Anomaly Detectioncited by 115
  9. 2024Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detectioncited by 93
  10. 2023Hierarchical Vector Quantized Transformer for Multi-class Unsupervised Anomaly Detectioncited by 66
  11. 2021A Hierarchical Transformation-Discriminating Generative Model for Few Shot Anomaly Detectioncited by 60
  12. 2020DFR: Deep Feature Reconstruction for Unsupervised Anomaly Segmentationcited by 54
  13. 2025Exploring Intrinsic Normal Prototypes within a Single Image for Universal Anomaly Detectioncited by 48
  14. 2023Focus the Discrepancy: Intra- and Inter-Correlation Learning for Image Anomaly Detectioncited by 32
  15. 2023ReConPatch : Contrastive Patch Representation Learning for Industrial Anomaly Detectioncited by 30
  16. 2020TrustMAE: A Noise-Resilient Defect Classification Framework using Memory-Augmented Auto-Encoders with Trust Regionscited by 21
  17. 2022A Feature Memory Rearrangement Network for Visual Inspection of Textured Surface Defects Toward Edge Intelligent Manufacturingcited by 16
  18. 2024AMI-Net: Adaptive Mask Inpainting Network for Industrial Anomaly Detection and Localizationcited by 12
  19. 2022Siamese Transition Masked Autoencoders as Uniform Unsupervised Visual Anomaly Detectorcited by 9
  20. 2022Normal Reference Attention and Defective Feature Perception Network for Surface Defect Detectioncited by 9
  21. 2024Dual-path Frequency Discriminators for Few-shot Anomaly Detectioncited by 7
  22. 2025A Masked Reverse Knowledge Distillation Method Incorporating Global and Local Information for Image Anomaly Detectioncited by 2

Built on by

Not cited yet.