[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-82365-en":3,"doc-seo-82365-105":29,"detail-sidebar-cat-0-en-105":95},{"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},82365,687197207919,"Theodora","https://ap-avatar.wpscdn.com/avatar/a000253d6f5f7c60be?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779446848396160552",8,"Research & Report","Decoupling Language Guidance from Backbones for Text-Guided Medical Segmentation","Text-guided medical image segmentation uses clinical semantics to improve lesion delineation, but many approaches tightly bind cross-modal fusion, supervision, and decoder design to specific architectures, making reuse difficult when vision or language backbones change. This paper introduces BTHA, a backbone-transferable hierarchical adapter framework with a stable feature-level interface and shape-preserving adapters. It employs hierarchical coarse-to-fine supervision and a scale-adaptive gated semantic guidance adapter, showing consistent effectiveness across convolutional and transformer visual encoders and multiple language encoders, with gains on four public datasets and modest overhead.","Decoupling Language Guidance from Backbones for Text-Guided Medical Segmentation  \nYungeng Liu∗ Harbin Institute of Technology (Shenzhen) Shenzhen, China [25B951006@stu.hit.edu.cn](25B951006@stu.hit.edu.cn)  \nXuanzi Fang∗ Harbin Institute of Technology (Shenzhen) Shenzhen, China [firetreehouse0@gmail.com](firetreehouse0@gmail.com)  \nHaijin Zeng  \nHarbin Institute of Technology (Shenzhen) Shenzhen, China [haijin.zeng2018@gmail.com](haijin.zeng2018@gmail.com)  \nQi Dai  \nNingBo No.2 Hospital NingBo, China [yxdaiqi@163.com](yxdaiqi@163.com)  \nYongyong Chen† Harbin Institute of Technology (Shenzhen) Shenzhen, China [cyy2020@hit.edu.cn](cyy2020@hit.edu.cn)  \narXiv :2607 .09481v1 [ cs .CV] 10 Jul 2026  \nAbstract—Text-guided medical image segmentation leverages clinical semantics to improve lesion delineation, yet many existing models bind cross-modal fusion, supervision, and decoder design into a task-specific architecture. Such tight coupling makes it difficult to reuse language guidance modules across heterogeneous vision and text backbones, and often requires redesigning the network when the encoder pair changes. This paper presents BTHA, a backbone-transferable hierarchical adapter framework for text-guided medical image segmentation. BTHA is built around a stable feature-level interface: given multi-scale visual features and a text representation, it injects semantic guidance through shape-preserving adapters while maintaining the decoder-side tensor contract. To make this interface effective, we introduce a Hierarchical Coarse-to-Fine Supervision Strategy that decomposes learning into global image-text alignment, multiscale auxiliary localization, and boundary-aware final mask refinement. We further design a Scale-Adaptive Gated Semantic Guidance (SAGSG) adapter, where resolution-specific gates adaptively control textual injection and channel recalibration suppresses redundant cross-modal responses. Evaluations across diverse vision and text backbones show that the same adapter and supervision design remains effective across convolutional and transformer-based visual encoders as well as different language encoders. Experiments on four public datasets further demonstrate that BTHA improves strong text-guided baselines with modest computational overhead.  \nIndex Terms—Medical Image Segmentation, Backbone Transferability, Vision-Language Models, Hierarchical Framework, Cross-Modal Alignment  \nI. INTRODUCTION  \nMedical image segmentation is a cornerstone of modern clinical analysis, supporting diagnosis, treatment planning, disease monitoring, and quantitative assessment [1] . Conventional automated segmentation methods mainly rely on visual appearance. Representative vision-only models, such as U-Net [2], nnU-Net [3], and UCTransNet [4], have advanced encoder– decoder design, self-configuring pipelines, transformer-based context modeling, and skip-connection fusion. However, be  \ncause these methods infer masks only from image appearance,∗ These authors contributed equally to this work.  \n† Corresponding author: Yongyong Chen.  \nthey remain vulnerable to low contrast, ambiguous lesion boundaries, anatomical variation, and limited pixel-level annotations [5], [6] . These challenges are particularly evident in lesion segmentation, where target regions can be small, diffuse, or visually similar to surrounding tissues. Clinical reports and textual descriptions provide complementary semantic cues, such as lesion type, anatomical location, and abnormality extent. Text-guided medical image segmentation therefore offers a natural way to use language as a semantic  \nprior for improving localization and delineation [7]–[9] .  \nRecent promptable segmentation models have further reshaped the segmentation landscape. The Segment Anything Model (SAM) [10] and its medical adaptations, including SAM-Adapter [11], MedSAM [12], and SAM3 [13], demonstrate impressive generalization through prompt-driven mask generation. However, these models usually d","cbCaiashjZ3wytax","https://ap.wps.com/l/cbCaiashjZ3wytax","pdf",2912960,1,9,"English","en",105,"# Abstract\n# Introduction\n## Motivation and clinical challenges\n## Related work in vision-only and promptable segmentation\n## Open problem: transferability of language guidance designs","[{\"question\":\"What problem does BTHA address in text-guided medical segmentation?\",\"answer\":\"BTHA targets the poor transferability of existing language-guidance designs, where fusion, supervision, and decoding are tightly coupled to specific vision-language backbones, requiring redesign when encoders change.\"},{\"question\":\"How does BTHA enable backbone transferability?\",\"answer\":\"BTHA uses a stable feature-level interface that injects semantic guidance through shape-preserving adapters while maintaining a consistent decoder-side tensor contract, enabling module reuse across different backbones.\"},{\"question\":\"What supervision and adaptation strategies does BTHA introduce?\",\"answer\":\"BTHA proposes a Hierarchical Coarse-to-Fine Supervision Strategy (global alignment, multiscale auxiliary localization, and boundary-aware refinement) and a Scale-Adaptive Gated Semantic Guidance (SAGSG) adapter with resolution-specific gating and channel recalibration.\"},{\"question\":\"What evidence supports BTHA’s effectiveness and efficiency?\",\"answer\":\"Experiments across diverse vision/text backbones and on four public datasets show that BTHA improves strong text-guided baselines while adding only modest computational overhead.\"}]",1784179932,23,{"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":90,"head_meta":92,"extra_data":94,"updated_unix":27},"decoupling-language-guidance-from-backbones-for-text-guided-medical-segmentation","",{"@graph":35,"@context":89},[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/decoupling-language-guidance-from-backbones-for-text-guided-medical-segmentation/82365/",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",[71,77,81,85],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does BTHA address in text-guided medical segmentation?","Question",{"text":75,"@type":76},"BTHA targets the poor transferability of existing language-guidance designs, where fusion, supervision, and decoding are tightly coupled to specific vision-language backbones, requiring redesign when encoders change.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does BTHA enable backbone transferability?",{"text":80,"@type":76},"BTHA uses a stable feature-level interface that injects semantic guidance through shape-preserving adapters while maintaining a consistent decoder-side tensor contract, enabling module reuse across different backbones.",{"name":82,"@type":73,"acceptedAnswer":83},"What supervision and adaptation strategies does BTHA introduce?",{"text":84,"@type":76},"BTHA proposes a Hierarchical Coarse-to-Fine Supervision Strategy (global alignment, multiscale auxiliary localization, and boundary-aware refinement) and a Scale-Adaptive Gated Semantic Guidance (SAGSG) adapter with resolution-specific gating and channel recalibration.",{"name":86,"@type":73,"acceptedAnswer":87},"What evidence supports BTHA’s effectiveness and efficiency?",{"text":88,"@type":76},"Experiments across diverse vision/text backbones and on four public datasets show that BTHA improves strong text-guided baselines while adding only modest computational overhead.","https://schema.org",{"og:url":51,"og:type":91,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":93,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":96},[97,101,105,109,114,119,124,127,131,134,138],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},"Exam",70,"exam",{"id":110,"doc_module":4,"doc_module_name":45,"category_name":111,"show_sort_weight":112,"slug":113},5,"Comic",60,"comic",{"id":115,"doc_module":4,"doc_module_name":45,"category_name":116,"show_sort_weight":117,"slug":118},6,"Technology",50,"technology",{"id":120,"doc_module":4,"doc_module_name":45,"category_name":121,"show_sort_weight":122,"slug":123},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":125,"slug":126},30,"research-report",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":129,"slug":130},"Religion & Spirituality",20,"religion-spirituality",{"id":129,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":129,"slug":133},"World Cup","world-cup",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":135,"slug":137},10,"Lifestyle","lifestyle",{"id":139,"doc_module":4,"doc_module_name":45,"category_name":140,"show_sort_weight":110,"slug":141},19,"General","general"]