UNI-OOD：通过跨上下文注意力视觉语言建模统一对象级与图像级分布外检测
UNI-OOD: Unified Object- and Image-level Out-of-Distribution Detection via Cross-Context Attentive Vision-Language Modeling
UNI-OOD 的跨上下文注意力机制可复用于任何需要细粒度对象级语义对齐的视觉语言任务，例如开放词表检测中的对象-背景关系建模；其统一的对象级与图像级 OOD 评分机制为视觉系统在开放环境下的可靠性评估提供了可直接借鉴的框架。
原文依据 We propose UNI-OOD, a unified framework that performs both object- and image-level OOD detection within a single VLM
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论分布外（OOD）检测是在开放世界环境中可靠部署的关键要求，模型必须识别出超出已知概念语义范围的输入。尽管视觉语言模型（VLM）的最新进展在图像级 OOD 检测上取得了强劲结果，但大多数方法仍假设每张图像包含单个主导对象。这一假设严重限制了它们在真实场景中的适用性，因为真实场景自然由多个对象组成，每个对象都需要独立的 OOD 评估。现有的对象级方法，包括当前最先进的 RUNA，仍受限于粗粒度的全局表示以及对对象与其背景之间上下文依赖建模不足。我们提出 UNI-OOD，一个统一的框架，在单个 VLM 内同时执行对象级和图像级 OOD 检测，且在推理时无需预先知道正在处理哪项任务。其核心思想是利用跨上下文注意力建模来捕获互补的视觉和文本语义。UNI-OOD 学习关注每个对象内部的细粒度空间细节，对齐视觉和语言嵌入以增强语义对应关系，并建模目标对象与其周围上下文之间的交互。通过对以对象为中心的线索和背景线索进行联合推理，该框架将信息性视觉证据与虚假关联解耦，并在不同视觉粒度上实现一致的 OOD 评分机制。在对象级和图像级基准上的大量实验表明，UNI-OOD 相较于先前方法取得了显著且一致的改进，在对象级和图像级 OOD 检测上均确立了新的最先进性能。
英文原文摘要 1818 字符
Out-of-distribution (OOD) detection is a key requirement for reliable deployment in open-world environments, where a model must recognize inputs that fall outside the semantic scope of known concepts. While recent advances in vision-language models (VLMs) have achieved strong results in image-level OOD detection, most methods still assume that each image contains a single dominant object. This assumption severely limits their applicability to real-world settings where scenes are naturally composed of multiple objects that each demand independent OOD assessment. Existing object-level approaches, including the current SOTA method RUNA, remain constrained by coarse global representations and insufficient modeling of contextual dependencies between objects and their backgrounds. We propose UNI-OOD, a unified framework that performs both object- and image-level OOD detection within a single VLM, without requiring prior knowledge of which task is being addressed at inference time. The key idea is to leverage cross-context attentive modeling that captures complementary visual and textual semantics. UNI-OOD learns to attend to fine-grained spatial details within each object, aligns visual and linguistic embeddings to strengthen semantic correspondence, and models interactions between target objects and their surrounding context. By jointly reasoning over object-centric and background cues, the framework disentangles informative visual evidence from spurious correlations and enables a consistent OOD scoring mechanism across different visual granularities. Extensive experiments on object- and image-level benchmarks demonstrate that UNI-OOD achieves substantial and consistent improvements over previous approaches, establishing new SOTA performance in both object-level and image-level OOD detection.