PWLR：用于边界感知分布外检测的成对见证局部拒绝
PWLR: Pairwise Witness Local Rejection for Boundary-Aware Out-of-Distribution Detection
浏览论文内容
中文总结 AI 辅助
针对视觉语言检测器极少用语言作为混淆ID类别边界证据的问题，提出PWLR方法，结合MLLM生成的局部线索与全局类别得分，在多基准上改进了视觉语言基线的OOD检测性能。
中文摘要 AI 辅助
分布外（OOD）检测对于图像分类器而言仍然具有挑战性，尤其是当近分布外（near-OOD）样本靠近分布内（ID）类别边界时。近期的视觉语言检测器通过类别语义、局部提示或大语言模型（LLM）生成的异常概念来改进OOD检测，但很少将语言作为混淆ID类别之间的显式边界证据。我们提出成对见证局部拒绝（PWLR），其使用多模态大语言模型（MLLM）离线描述有利于某一ID类别优于特定竞争类别的可见局部线索。随后，这些线索短语在仅含ID数据的情况下，通过冻结的视觉语言骨干网络进行筛选，仅保留可靠的局部验证器。推理时，PWLR首先保留一小部分全局合理的类别，然后检查其中是否有任何类别在局部上得到其最相关竞争类别的支持，最后通过校准将此成对局部证据与全局类别得分结合。在ImageNet-100远OOD、更清洁/具挑战性的OOD以及近OOD基准上的实验表明，PWLR在多个骨干网络上均能持续改进强大的视觉语言基线。源代码将发布。
英文摘要
Out-of-distribution (OOD) detection remains challenging for image classifiers, especially when near-OOD samples lie close to in-distribution (ID) class boundaries. Recent vision-language detectors improve OOD detection through class semantics, local prompting, or LLM-generated outlier concepts, but seldom use language as explicit boundary evidence between confusing ID classes. We propose Pairwise Witness Local Rejection (PWLR), which uses an MLLM offline to describe visible local cues that favor one ID class over a specific rival class. These cue phrases are then screened with ID-only data under a frozen vision-language backbone, so that only reliable local verifiers are kept. At inference, PWLR first retains a small set of globally plausible classes, then checks whether any of them is locally supported against its most relevant rivals, and finally combines this pairwise local evidence with the global class score through calibration. Experiments on ImageNet-100 far-OOD, cleaner/challenging OOD and near-OOD benchmarks show that PWLR consistently improves strong vision-language baselines across multiple backbones. Source code will be released.
发表机构
- Guangdong Provincial Key Laboratory of Stomatology（广东省口腔医学重点实验室）
- Key Laboratory of Machine Intelligence and Advanced Computing, MOE（教育部机器智能与先进计算重点实验室）
机构由 AI 辅助整理，请以论文原文为准。
