基于推理驱动的图像级监督异常检测与定位
Reasoning-Driven Anomaly Detection and Localization with Image-Level Supervision
利用MLLM自回归推理过程中的注意力响应生成像素级异常图，仅需图像级监督，为视觉异常检测提供无需密集标注的新范式。
原文依据 extracts anomaly-related tokens from the autoregressive reasoning process and aggregates their attention responses to produce pixel-level anomaly maps
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论多模态大语言模型（MLLMs）近期在异常检测方面展现出卓越的推理与感知能力。然而，大多数方法仍局限于图像级异常检测和文本推理，而像素级定位仍依赖外部视觉模块和密集标注。在本工作中，我们激活MLLMs的内在推理潜力，仅依靠图像级监督即可执行异常检测、像素级定位和可解释推理，无需任何辅助组件或像素级标签。具体而言，我们提出推理驱动异常定位（ReAL），该方法从自回归推理过程中提取异常相关token，并聚合其注意力响应以生成像素级异常图。我们进一步引入一致性引导推理优化（CGRO）模块，利用强化学习使推理token与视觉注意力对齐，从而产生更连贯的推理和更准确的异常定位。在四个公开基准上的大量实验表明，我们的方法显著提升了异常检测、定位和可解释性。值得注意的是，尽管仅依赖图像级监督，我们的方法取得了与在密集像素级监督下训练的基于MLLM的方法相竞争的性能。代码见 https://github.com/YizhouJin313/ReADL。
英文原文摘要 1432 字符
Multimodal large language models (MLLMs) have recently demonstrated remarkable reasoning and perceptual abilities for anomaly detection. However, most approaches remain confined to image-level anomaly detection and textual reasoning, while pixel-level localization still relies on external vision modules and dense annotations. In this work, we activate the intrinsic reasoning potential of MLLMs to perform anomaly detection, pixel-level localization, and interpretable reasoning solely from image-level supervision, without any auxiliary components or pixel-wise labels. Specifically, we propose Reasoning-Driven Anomaly Localization (ReAL), which extracts anomaly-related tokens from the autoregressive reasoning process and aggregates their attention responses to produce pixel-level anomaly maps. We further introduce a Consistency-Guided Reasoning Optimization (CGRO) module that leverages reinforcement learning to align reasoning tokens with visual attentions, resulting in more coherent reasoning and accurate anomaly localization. Extensive experiments on four public benchmarks demonstrate that our method significantly improves anomaly detection, localization, and interpretability. Remarkably, despite relying solely on image-level supervision, our approach achieves performance competitive with MLLM-based methods trained under dense pixel-level supervision. Code is available at https://github.com/YizhouJin313/ReADL.