生成、分析与精修：基于多模态大模型元推理的免训练声源定位
Generate, Analyze, and Refine: Training-Free Sound Source Localization via MLLM Meta-Reasoning
把 MLLM 的推理能力当作免训练的音视频一致性判别器：用开放集角色标注加锚框投票量化音视频一致性，再用自适应门控决定是否修正框，这套「生成-分析-精修」流程可直接迁移到其他需要跨模态验证的定位任务（如指代定位、开放词表检测的后处理）。
原文依据 Analysis quantifies Audio-Visual Consistency via open-set role tagging and anchor voting
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论声源定位任务旨在利用音频与视觉模态之间的相关性来识别发声物体的位置。现有的大多数声源定位方法依赖基于对比学习的特征匹配，但缺乏显式的推理与验证，限制了其在复杂声学场景中的有效性。受人类元认知过程的启发，我们提出了一种免训练的声源定位框架，充分利用多模态大语言模型（MLLM）内在的推理能力。我们的「生成-分析-精修」（GAR）流程包含三个阶段：生成阶段产生初始边界框与音频分类；分析阶段通过开放集角色标注与锚框投票量化音视频一致性；精修阶段施加自适应门控以避免不必要的调整。在单声源与多声源基准上的大量实验表明该方法具有竞争力的性能。源代码见 https://github.com/VisualAIKHU/GAR-SSL。
英文原文摘要 1001 字符
Sound source localization task aims to identify the locations of sound-emitting objects by leveraging correlations between audio and visual modalities. Most existing SSL methods rely on contrastive learning-based feature matching, but lack explicit reasoning and verification, limiting their effectiveness in complex acoustic scenes. Inspired by human meta-cognitive processes, we propose a training-free SSL framework that exploits the intrinsic reasoning capabilities of Multimodal Large Language Models (MLLMs). Our Generation-Analysis-Refinement (GAR) pipeline consists of three stages: Generation produces initial bounding boxes and audio classifications; Analysis quantifies Audio-Visual Consistency via open-set role tagging and anchor voting; and Refinement applies adaptive gating to prevent unnecessary adjustments. Extensive experiments on single-source and multi-source benchmarks demonstrate competitive performance. The source code is available at https://github.com/VisualAIKHU/GAR-SSL.