mlsanityData to 14 Oct 2026

UniVAD v2: Unified Visual Anomaly Detection via Support-Conditioned Boundary Construction

arXiv 2606.29714 ↗First posted 29 Jun 2026cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Unified visual anomaly detection seeks to train a single detector that can be deployed across categories, domains, and application scenarios. In the few-shot transfer regime, the key challenge is to estimate an episode-specific boundary for an unseen target category from a small support set. Existing approaches mainly infer this boundary from normal-side evidence and provide limited abnormal-side evidence for deployment-specific tolerance. Within the normal side, they often struggle to jointly capture local correspondences and global support-query relations, making their boundaries less reliable for unseen anomalies. To address these issues, we propose UniVAD v2, a two-sided support-conditioned boundary construction framework for unified visual anomaly detection. Built on the component-patch divide-and-conquer framework of UniVAD, UniVAD v2 strengthens the normal side with an Optimal Transport-based Relational Modeling module (OTRM), which complements retrieval with support-query matching through transport-style allocation, and an Adaptive Coordination mechanism for Retrieval and Relational Modeling (ACRRM), which estimates episode-conditioned reliabilities to fuse the two sources of evidence. On the abnormal side, a Few-Shot Abnormal Reference module (FAR) converts optional abnormal references into rejection-side evidence for boundary adjustment. Experiments on six datasets spanning industrial, logical, and medical anomaly detection demonstrate strong cross-domain generalization. Under the 1N-shot protocol, UniVAD v2 improves the mean image-level AUC over UniVAD from 83.0% to 84.5%, and further reaches 85.7% in the 1N+1A-shot setting. On the MVTec-AD Severity Split (MVTec-AD-SS), UniVAD v2 achieves 96.2% image-level AUC and 96.9% pixel-level AUC, showing that abnormal references enable controllable boundary customization without retraining.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 35 of its 54 references that are in the map. First posted on arXiv 29 Jun 2026.

Builds on

35 of 54 references.

References · 35

  1. 2017Attention Is All You Needcited by 60,361
  2. 2017Deep Setscited by 1,421
  3. 2021Towards Total Recall in Industrial Anomaly Detectioncited by 776
  4. 2020PaDiM: a Patch Distribution Modeling Framework for Anomaly Detection and Localizationcited by 594
  5. 2022MedCLIP: Contrastive Learning from Unpaired Medical Images and Textcited by 570
  6. 2021The RSNA-ASNR-MICCAI BraTS 2021 Benchmark on Brain Tumor Segmentation and Radiogenomic Classificationcited by 518
  7. 2022Anomaly Detection via Reverse Distillation from One-Class Embeddingcited by 421
  8. 2022SPot-the-Difference Self-Supervised Pre-training for Anomaly Detection and Segmentationcited by 400
  9. 2021CFLOW-AD: Real-Time Unsupervised Anomaly Detection with Localization via Conditional Normalizing Flowscited by 321
  10. 2023WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentationcited by 315
  11. 2020Sub-Image Anomaly Detection with Deep Pyramid Correspondencescited by 300
  12. 2022A Unified Model for Multi-class Anomaly Detectioncited by 268
23 more
  1. 2020Anomaly Detection in Video via Self-Supervised and Multi-Task Learningcited by 148
  2. 2022Registration based Few-Shot Anomaly Detectioncited by 133
  3. 2024Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detectioncited by 93
  4. 2023BMAD: Benchmarks for Medical Anomaly Detectioncited by 82
  5. 2024Adapting Visual-Language Models for Generalizable Anomaly Detection in Medical Imagescited by 69
  6. 2023Open-Vocabulary Video Anomaly Detectioncited by 57
  7. 2024FiLo: Zero-Shot Anomaly Detection by Fine-Grained Description and High-Quality Localizationcited by 51
  8. 2022Self-Supervised Masked Convolutional Transformer Block for Anomaly Detectioncited by 34
  9. 2024UniVAD: A Training-free Unified Model for Few-shot Visual Anomaly Detectioncited by 28
  10. 2024Transformers Can Learn Temporal Difference Methods for In-Context Reinforcement Learningcited by 23
  11. 2023Component-aware anomaly detection framework for adjustable and logical industrial visual inspectioncited by 22
  12. 2025DictAS: A Framework for Class-Generalizable Few-Shot Anomaly Segmentation via Dictionary Lookupcited by 10
  13. 2025Normal-Abnormal Guided Generalist Anomaly Detectioncited by 10
  14. 2021Bayesian Embeddings for Few-Shot Open World Recognitioncited by 8
  15. 2025FiLo++: Zero-/Few-Shot Anomaly Detection by Fused Fine-Grained Descriptions and Deformable Localizationcited by 8
  16. 2025AnomalyMoE: Towards a Language-free Generalist Model for Unified Visual Anomaly Detectioncited by 8
  17. 2025Robust Distribution Alignment for Industrial Anomaly Detection under Distribution Shiftcited by 5
  18. 2025LOTFormer: Doubly-Stochastic Linear Attention via Low-Rank Optimal Transportcited by 4
  19. 2023Learnable Graph Matching: A Practical Paradigm for Data Associationcited by 2
  20. 2025Quality-Aware Language-Conditioned Local Auto-Regressive Anomaly Synthesis and Detectioncited by 2
  21. 2026RAID: Retrieval-Augmented Anomaly Detectioncited by 2
  22. 2024Prototype-based Optimal Transport for Out-of-Distribution Detectioncited by 1
  23. 2024Provable optimal transport with transformers: The essence of depth and prompt engineeringcited by 1

Built on by

Not cited yet.