mlsanityData to 14 Oct 2026

What Memory Composition Does Not Tell Us About Anomaly Detection

arXiv 2608.23295 ↗First posted 24 Aug 2026cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Memory-based anomaly detectors store nominal training patches and score test patches against this memory. A patch selected for coverage therefore becomes a nor- mal reference without a separate check that geometric rarity makes it safe to trust. We probe this coupling with sparse training contamination. Under fixed representa- tions and memory budgets, we compare random, medoid, local, and global coverage selectors. We then use CLEANCON, an out-of-bag cross-image support gate that changes candidate-image eligibility while fixing the representation, absolute mem- ory size, builder, and inference rule. Global coverage strongly over-represents sparse contamination. CLEANCON reduces final-memory contamination to approx- imately zero and increases category-macro P-AP in all 12 matched comparisons. Yet along a retention sweep, the lowest-contamination memory does not attain the highest P-AP; performance continues to improve while contamination rises. Mem- ory contamination therefore does not order the resulting memories by P-AP

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 18 of its 30 references that are in the map. First posted on arXiv 24 Aug 2026.

Builds on

18 of 30 references.

References · 18

  1. 2021Towards Total Recall in Industrial Anomaly Detectioncited by 776
  2. 2020PaDiM: a Patch Distribution Modeling Framework for Anomaly Detection and Localizationcited by 594
  3. 2022SPot-the-Difference Self-Supervised Pre-training for Anomaly Detection and Segmentationcited by 400
  4. 2020Sub-Image Anomaly Detection with Deep Pyramid Correspondencescited by 300
  5. 2021GRAD-MATCH: Gradient Matching based Data Subset Selection for Efficient Deep Model Trainingcited by 274
  6. 2020Coresets via Bilevel Optimization for Continual Learning and Streamingcited by 192
  7. 2020PANDA: Adapting Pretrained Features for Anomaly Detection and Segmentationcited by 175
  8. 2021Mean-Shifted Contrastive Loss for Anomaly Detectioncited by 88
  9. 2024SoftPatch: Unsupervised Anomaly Detection with Noisy Datacited by 55
  10. 2022Latent Outlier Exposure for Anomaly Detection with Contaminated Datacited by 48
  11. 2017Constant Approximation for k-Median and k-Means with Outliers via Iterative Roundingcited by 25
  12. 2021Self-supervise, Refine, Repeat: Improving Unsupervised Anomaly Detectioncited by 20
6 more
  1. 2024FUN-AD: Fully Unsupervised Learning for Anomaly Detection with Noisy Training Datacited by 8
  2. 2024SoftPatch+: Fully Unsupervised Anomaly Classification and Segmentationcited by 7
  3. 2022Data refinement for fully unsupervised visual inspection using pre-trained networkscited by 4
  4. 2025A Synergy Scoring Filter for Unsupervised Anomaly Detection with Noisy Datacited by 1
  5. 2025Leveraging Learning Bias for Noisy Anomaly Detectioncited by 1
  6. 2026Memory-Distilled Selection for Noise-Robust Anomaly Detectioncited by 1

Built on by

Not cited yet.