从网络到像素：将代理搜索引入视觉感知
From Web to Pixels: Bringing Agentic Search into Visual Perception
AI总结：
本文提出感知深度研究，引入WebEye基准，通过代理搜索到像素的工作流程，解决隐藏目标身份并绑定到框、掩码或 grounded 答案，实验显示其在三个任务视图中表现最佳。
摘要：
视觉感知连接高层语义理解到像素级感知，但现有方法大多假设目标识别的决定性证据已存在于图像或冻结模型知识中。本文研究了更实际但更难的开放世界场景，其中可见对象必须首先从外部事实、近期事件、长尾实体或多跳关系中解析后才能定位。我们正式将这一挑战定义为感知深度研究，并引入WebEye基准，该基准包含可验证的证据、知识密集型查询、精确的框/掩码注释以及三个任务视图：基于搜索的 grounding、基于搜索的分割和基于搜索的 VQA。WebEyes 包含 120 张图像、473 个注释对象实例、645 个唯一 QA 对和 1,927 个任务样本。我们进一步提出 Pixel-Searcher，一种代理搜索到像素的工作流程，能够解析隐藏目标身份并将其绑定到框、掩码或 grounded 答案。实验显示，Pixel-Searcher 在所有三个任务视图中实现了最强的开源性能，而失败主要源于证据获取、身份解析和视觉实例绑定。
英文摘要：
Visual perception connects high-level semantic understanding to pixel-level perception, but most existing settings assume that the decisive evidence for identifying a target is already in the image or frozen model knowledge. We study a more practical yet harder open-world case where a visible object must first be resolved from external facts, recent events, long-tail entities, or multi-hop relations before it can be localized. We formalize this challenge as Perception Deep Research and introduce WebEye, an object-anchored benchmark with verifiable evidence, knowledge-intensive queries, precise box/mask annotations, and three task views: Search-based Grounding, Search-based Segmentation, and Search-based VQA. WebEyes contains 120 images, 473 annotated object instances, 645 unique QA pairs, and 1,927 task samples. We further propose Pixel-Searcher, an agentic search-to-pixel workflow that resolves hidden target identities and binds them to boxes, masks, or grounded answers. Experiments show that Pixel-Searcher achieves the strongest open-source performance across all three task views, while failures mainly arise from evidence acquisition, identity resolution, and visual instance binding.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
