[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84531-en":3,"doc-seo-84531-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},84531,962075006959,"Anda","https://ap-avatar.wpscdn.com/avatar/e0002397efbe92a78e?_k=1776741047341049297",8,"Research & Report","PixelEyes: Decoupling Perception and Reasoning for Pinpoint Visual Evidence Seeking","The paper studies multi-turn visual reasoning in which multimodal large language models frequently fail to localize the target, generating long redundant trajectories. The work attributes this to entanglement between perception and reasoning: the model reasons and grounds in a single loop, and inaccurate localization triggers extra turns that inflate search paths. PixelEyes addresses this by decoupling reasoning from perception via a mask-guided tool and a semantic-region BFS exploration strategy, with dedicated benchmarks and open-sourced code and models.","arXiv :2607 .00115v1 [ cs .CV] 30 Jun 2026  \nPixelEyes: Decoupling Perception and Reasoning for Pinpoint Visual Evidence Seeking  \nDengxian Gong 1 * Yuanzheng Wu 1 * Haobo Yuan2 Zhengdong Hu3 Tao Zhang 1 Yikang Zhou 1 Shihao Chen 1 Quanzhu Niu 1 Kai Wang4 Jason Li5 Haochen Wang6 Lu Qi 1† Shunping Ji 1† Ming-Hsuan Yang2  \n1Wuhan University 2UC Merced 3UTS 4NUS 5NTU 6 CASIA  \n*Equal contribution, †Corresponding authors  \n{gooodx,[jishunping}@whu.edu.cn](jishunping}@whu.edu.cn)[ ](jishunping}@whu.edu.cn)Project page: [https://godx-7.github.io/PixelEyesSite/](https://godx-7.github.io/PixelEyesSite/)  \nFigure 1: Paradigm Comparison for Active Visual Search. (a) A challenging instance-anchored visual query. (b) Coupled Agent (Baseline): Relying on coarse bounding boxes, existing models suffer from \"inattentional blindness\" (spotting the correct region but failing to recognize the target) and fall into rigid, inefficient deep-search loops, eventually exhausting the turn limit. (c) Decoupled Agent (PixelEyes): By employing the SAMTok tool for precise, mask-guided cropping and adopting a Semantic-Region Breadth-First Search (BFS) exploration strategy, our framework eliminates background distractors and efficiently locates the target in just 5 turns. (d) Consequently, PixelEyes achieves state-of-the-art performance across multiple rigorous visual search benchmarks, significantly outperforming existing methods.  \nAbstract  \nThis paper explores multi-turn visual reasoning and observes that MLLMs repeatedly fail to localize the target, leading to long, redundant trajectories. We attribute this failure to the entanglement of reasoning and perception within a single model, the MLLM reasons and localizes simultaneously, and inaccurate localization triggers additional reasoning turns that bloat the trajectory. To solve this problem, we  \nPreprint.  \npropose PixelEyes, a multi-turn visual reasoning agent that explicitly decouples reasoning from perception, i.e., the reasoner decides what to look for, while a specialized perception tool answers where it is. Specifically, PixelEyes introduces  \n1) Mask-guided Visual Search. A referring segmentation model is invoked to provide mask-precise localization, freeing the reasoner from the need to compensate for imprecise grounding. 2) Semantic-region Breadth-first Search (BFS) . To eliminate redundant loops caused by repeatedly cropping incorrect sub-regions, we organize exploration as a breadth-first search over semantic regions. To internalize these capabilities, we construct the PixelEyes-6K dataset by resynthesizing expert trajectories from existing data. This explicitly embeds our mask-guided search and BFS logic into the model. We further introduce Pinpoint-Bench, a zero-hint visual search benchmark, i.e., no location cues are provided in the question, with instance-level masks and bounding boxes that separate localization failures from reasoning failures, enabling fine-grained analysis of failure modes such as inattentional blindness. Recent state-of-the-art MLLMs and visual reasoning agents leave large headroom on Pinpoint-Bench, demonstrating its quality and difficulty. Code and models are open-sourced.  \n1 Introduction  \nVision-language models are moving from passive observers to active visual reasoners that crop, zoom, and re-examine an image to gather evidence – a paradigm popularized by OpenAI o3 [22] and now widely studied as “Thinking with Images” [26] . The setting is challenging in practice: decisive evidence often lies in objects occupying less than 1% of a high-resolution image, so an agent must locate a needle-in-a-haystack target and then reason about its content within a bounded turn budget. Existing methods [25, 40, 51, 14, 24, 18] pursue this with a single model that is asked to perform both jobs at once: fine-grained region-level perception and general reasoning over the cropped evidence. This coupling is uncomfortable. The same model is consistently weaker at grounding than per","cbCaire6d14ZIYtR","https://ap.wps.com/l/cbCaire6d14ZIYtR","pdf",25711424,1,22,"English","en",105,"# Introduction\n## PixelEyes motivation and decoupled agent design\n# Mask-guided Visual Search\n## Referring segmentation for precise localization\n# Semantic-region Breadth-first Search (BFS)\n## Reducing redundant loops in exploration","[{\"question\":\"What problem does PixelEyes address in multi-turn visual reasoning?\",\"answer\":\"MLLM-based agents often fail to localize the target, which causes long redundant trajectories and leads to ineffective exploration within a limited turn budget.\"},{\"question\":\"How does PixelEyes decouple perception and reasoning?\",\"answer\":\"A general-purpose VLM decides what to look for, while an external referring-segmentation tool (SAMTok) provides precise pixel-level masks indicating where the target is.\"},{\"question\":\"Why does PixelEyes use semantic-region BFS instead of a simple deep search loop?\",\"answer\":\"Semantic-region BFS organizes exploration to expand sibling semantic regions when grounding fails, avoiding repeated cropping of incorrect sub-regions and keeping trajectories short.\"}]",1784196456,55,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"pixeleyes-decoupling-perception-and-reasoning-for-pinpoint-visual-evidence-seeking","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/pixeleyes-decoupling-perception-and-reasoning-for-pinpoint-visual-evidence-seeking/84531/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does PixelEyes address in multi-turn visual reasoning?","Question",{"text":75,"@type":76},"MLLM-based agents often fail to localize the target, which causes long redundant trajectories and leads to ineffective exploration within a limited turn budget.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does PixelEyes decouple perception and reasoning?",{"text":80,"@type":76},"A general-purpose VLM decides what to look for, while an external referring-segmentation tool (SAMTok) provides precise pixel-level masks indicating where the target is.",{"name":82,"@type":73,"acceptedAnswer":83},"Why does PixelEyes use semantic-region BFS instead of a simple deep search loop?",{"text":84,"@type":76},"Semantic-region BFS organizes exploration to expand sibling semantic regions when grounding fails, avoiding repeated cropping of incorrect sub-regions and keeping trajectories short.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":45,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]