[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-327244-105":59,"doc-detail-327244-en":130},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":123,"head_meta":125,"extra_data":127,"updated_unix":129},105,"en","geoselect-spatial-program-execution-for-training-free-referring-remote-sensing-image-segmentation","GeoSelect - Spatial-Program Execution for Training-Free Referring Remote Sensing Image Segmentation","","Referring remote sensing image segmentation assigns an object to a natural-language expression and produces precise instance masks in aerial imagery. Prior training-free approaches infer expressions via implicit vision–language activations or region–text similarity, but offer limited control over spatial, comparative, and ordinal relations such as “largest ship” or “second court from the left.” GeoSelect reframes referring as typed spatial-program execution using a frozen text-only language model, a wellformedness checker, and a deterministic executor that unifies continuous geometric fields with discrete set and order operators. The method reaches 58.86 mIoU on RRSIS-D and 55.27 mIoU on RISBench without referring supervision on a single GPU.",{"@graph":69,"@context":122},[70,84,105],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/geoselect-spatial-program-execution-for-training-free-referring-remote-sensing-image-segmentation/327244/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/geoselect-spatial-program-execution-for-training-free-referring-remote-sensing-image-segmentation/327244.png","ImageObject",300,407,{"name":92,"@type":93},"River Wang","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-11-07","2026-09-21",true,{"@type":102,"interactionType":103,"userInteractionCount":34},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"What problem does GeoSelect address in referring remote sensing image segmentation?","Question",{"text":112,"@type":113},"It addresses weak control of spatial, comparative, and ordinal relations when mapping natural-language expressions to the correct instance in aerial scenes.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"How does GeoSelect differ from prior training-free approaches?",{"text":117,"@type":113},"GeoSelect executes a synthesized typed spatial program deterministically, rather than relying only on implicit vision–language activations or region–text similarity.",{"name":119,"@type":110,"acceptedAnswer":120},"What components does GeoSelect use to turn an expression into a segmentation result?",{"text":121,"@type":113},"A frozen text-only language model synthesizes a program in a small domain-specific language, a wellformedness checker validates it, and a deterministic executor runs it to produce results on a scored candidate set type.","https://schema.org",{"og:url":83,"og:type":124,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":126,"canonical":83},"index,follow",{"doc_id":128,"site_id":62},327244,1790668020,{"code":4,"msg":5,"data":131},{"doc_id":128,"user_id":132,"nickname":92,"user_avatar":133,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":34,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":46,"language":139,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":140,"faqs":141,"seo_title":142,"seo_description":67,"update_tm":143,"read_time":31,"preview_page_count":4},1099514067438,"https://ap-avatar.wpscdn.com/avatar/100002539ee87300030?x-image-process=image/resize,m_fixed,w_180,h_180&k=1780474512215547542","GeoSelect: Spatial-Program Execution for Training-Free Referring Remote Sensing Image Segmentation  \nYuhang Jiang, Guohui Deng, Miaozhong Xu, Chao Ruan, Jinling Zhao, and Linsheng Huang  \narXiv :2607 .03869v 1 [ cs .CV] 4 Jul 2026  \nAbstract—Referring remote sensing image segmentation identifies and segments the object named by a natural-language expression in an aerial image. Existing training-free methods resolve the expression through implicit vision–language activations or region–text similarity, which gives weak control over the spatial, comparative, and ordinal relations that dominate aerial referring: they cannot represent constructions such as the largest ship orthe second court from the left. We propose GeoSelect, a trainingfree pipeline that reframes referring as the execution of a typed spatial program. A frozen, text-only language model synthesises the expression into a small domain-specific language, a wellformedness checker accepts the program, and a deterministic executor runs it. The central abstraction is a single scored candidate set type under which every operator composes: continuous geometric fields realise position and proximity as dense pixel-level maps, while discrete set and order operators add the extremum, ordinal, counted-union, and relational constructions that fields alone cannot express. Because execution is explicit, every intermediate program, field, and ranking is inspectable, and a reliability ladder degrades any failing program to the field-only special case, so every expression still returns an answer. GeoSelect attains 58.86 mIoU on RRSIS-D test and 55.27 mIoU on RISBench test, more than twice the best prior training-free method on RRSIS-D, with no referring supervision and on a single GPU. A controlled comparison with candidates and the segmenter held fixed attributes the gain to explicit execution rather than the backbone; an oracle decomposition attributes the residual gap to detection recall on RRSIS-D and to candidateselection on RISBench, and an image-level exposure audit on RRSIS-D confirms robustness to detector-pretraining leakage. Code and configurations will be released upon acceptance atthe project page [https://avalon-s.github.io/GeoSelect/](https://avalon-s.github.io/GeoSelect/).  \nIndex Terms—Referring image segmentation, remote sensing, training-free, neuro-symbolic reasoning, program synthesis, vision–language models, spatial reasoning.  \nI. INTRODUCTION  \nREFERRING  \n(RRSIS) [1]  \nremote sensing image segmentation requires instance-level disambiguation in  \ncrowded aerial scenes. Expressions frequently specify the referent through spatial, comparative, and ordinal cues such asthe upper-right ship, the largest tank, or the second court from the left, where several same-class objects would otherwise be valid candidates. Resolving the referent therefore demands that the relational content of the expression be evaluated, not merely that the object category be recognised.  \nYuhang Jiang, Guohui Deng, Chao Ruan, Jinling Zhao, and Linsheng Huang are with the School of Internet, Anhui University, Hefei, China. Guohui Deng is also with the National Engineering Research Center for Agro-Ecological Big Data Analysis & Application, Anhui University, Hefei, China.  \nMiaozhong Xu is with the State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan, China.  \nCorresponding author: Guohui Deng (e-mail: [dlyj0207@whu.edu.cn](dlyj0207@whu.edu.cn)).  \nRecent supervised networks [1]–[3] reach strong accuracy on RRSIS-D, but require dense referring masks and a separate training run for each benchmark. Training-free methods instead combine frozen pretrained components, yet the best published result on RRSIS-D test remains far below the supervised specialists. The gap suggests that pretrained perception alone does not solve spatial referring.  \nWe study the representation of the spatial constraint (Fig. 1) . Existing training-free pipelines","cbCaisktGYHsDOgk","https://ap.wps.com/l/cbCaisktGYHsDOgk","pdf",11587478,"English","# Abstract\n# I. Introduction\n## Referring remote sensing image segmentation (RRSIS)\n## Limits of existing training-free methods\n## GeoSelect as spatial-program execution","[{\"question\":\"What problem does GeoSelect address in referring remote sensing image segmentation?\",\"answer\":\"It addresses weak control of spatial, comparative, and ordinal relations when mapping natural-language expressions to the correct instance in aerial scenes.\"},{\"question\":\"How does GeoSelect differ from prior training-free approaches?\",\"answer\":\"GeoSelect executes a synthesized typed spatial program deterministically, rather than relying only on implicit vision–language activations or region–text similarity.\"},{\"question\":\"What components does GeoSelect use to turn an expression into a segmentation result?\",\"answer\":\"A frozen text-only language model synthesizes a program in a small domain-specific language, a wellformedness checker validates it, and a deterministic executor runs it to produce results on a scored candidate set type.\"}]","GeoSelect - Spatial-Program Execution for Training-Free Referring Remote Sensing Image Segmentation | PDF",1789976798]