UniVAD v2: Unified Visual Anomaly Detection via Support-Conditioned Boundary Construction
2026-06-29 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors work on making a single tool that can spot unusual things (anomalies) in images from many different categories and situations, even with only a few examples to learn from. They improve their earlier method by better understanding both normal and abnormal examples to draw clearer boundaries between usual and unusual stuff. Their new approach uses smarter techniques to compare parts of images and adjusts boundaries using extra abnormal examples when available. Tests on various datasets show their method works well across different fields like industry and medicine. Overall, their system detects anomalies more accurately without needing retraining for new situations.
visual anomaly detectionfew-shot learningoptimal transportsupport-query matchingcross-domain generalizationabnormal referenceimage-level AUCpixel-level AUCboundary construction
Authors
Zhaopeng Gu, Bingke Zhu, Zhaowen Li, Guibo Zhu, Yingying Chen, Ming Tang, Peng Su, Jinqiao Wang
Abstract
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.