Newest papers

Agentic Ensemble with Vision-Language Models for Visual Anomaly Detection

Shangquan Sun, Wenjun Luo, Xu He, Yoonsoo Lee, Shijian Lu · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.28560 · cited by 0 works (OpenAlex) · open-access PDF

Korean

MVUR(Visual Transfer Anomaly Correction)는 비정상적인 개체를 감지하는 것을 목표로 하고 있으며, 비정상적인 샘플은 모델 훈련 중에 종종 부족합니다. 여러 개의 프리 트레이닝된 디테이커를 조립하면 전송 가능한 이미지 점수와 비정상적인 지도를 활용할 수 있습니다. 그러나 대부분의 기존 집합 방법은 저장된 예측에 대한 고정 융합 규칙을 적용하며, 프리 트레이닝된 디테이커가 다른 훈련 목표, 프리어 및 시각적 표현 때문에 명확하게 동의하지 않을 때 투쟁합니다. 이 목적을 위해, 우리는 AnomalyAgent, 정확한 시각적 비정상 감지를 위한 포스트 트레이닝 융합 프레임 워크를 설계합니다. BenomalyAgent는

Machine translation (Meta M2M100); the English original follows.

Abstract

Visual anomaly detection aims to detect abnormal objects, while anomaly samples are often scarce during model training. Assembling multiple pretrained detectors can help by leveraging their transferable image scores and anomaly maps. However, most existing ensemble methods apply fixed fusion rules over saved predictions which struggle when pretrained detectors clearly disagree due to different training objectives, priors, and visual representations. To this end, we design AnomalyAgent, a post-training fusion framework for accurate visual anomaly detection. AnomalyAgent aligns heterogeneous scores and maps, summarizes score- and region-level disagreement, and treats fusion as a finite action-selection problem. It implements two types of actions: 1) score actions that determine how to fuse image-level detector confidence via consensus, robust aggregation, detector weighting, or score-map consistency; 2) map actions that fuse dense anomaly responses via consensus, detector weighting, geometric aggregation, or disagreement correction. A constrained vision-language model (VLM) is introduced to inspect high-conflict samples and select valid score action and map action to produce the final prediction. We evaluate AnomalyAgent on MVTec AD and VisA as primary benchmarks, together with seven additional public anomaly benchmarks and Real-IAD for large-scale score-level evaluation. Across these benchmarks, AnomalyAgent consistently improves over equal-weight fusion and achieves the best mean image AUROC and pixel AP among the compared fusion methods on MVTec AD and VisA, while demonstrating robust transfer across additional anomaly domains.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Fuse (electrical), Anomaly detection, Anomaly (physics), Artificial intelligence, Detector, Computer science

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange