[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-81506-en":3,"doc-seo-81506-105":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},81506,1099513958762,"Logic","https://ap-avatar.wpscdn.com/avatar/1000023916a998db790?x-image-process=image/resize,m_fixed,w_180,h_180&k=1784791008015729253",8,"Research & Report","Prototypical Few-Shot Medical Image Semantic Segmentation with Background Fusion","Few-shot semantic segmentation (FSS) adapts a pretrained model to novel classes using only one labeled support sample per class. Prior prototypical methods emphasize foreground discrimination while simplifying background representation, assuming clear foreground–background separation common in natural images. Frequency spectrum entropy analysis shows medical foreground and background share many visual features, requiring richer background modeling. This paper proposes Background-fused prototype (Bro) with FeaC and HiCA modules that calibrate support noise and fuse background into discriminative prototypes, improving prior FSS performance on benchmarks.","arXiv :2412 .02983v2 [ cs .CV] 10 Jul 2026  \nPrototypical Few-Shot Medical Image Semantic Segmentation with Background Fusion  \nYuan Donga , Xiaoyu Yub , Wentao Wana , Jianchao Xuea , Yuejin Duanb , Song  \nTangb,∗, Yu Zhaob,∗  \na Chinese Academy of Medical Sciences and Peking Union Medical College, Department of  \nOrthopedics, Peking Union Medical College Hospital, Beijing, China b School of Health Sciences and Engineering, University of Shanghai for Science and  \nTechnology, Shanghai, China  \nAbstract  \nFew-shot Semantic Segmentation (FSS) aims to adapt a pre-trained model to new classes with as few as a single labeled training sample per class. The existing prototypical work used in natural image scenarios biasedly focus on capturing foreground’s discrimination while employing a simplistic representation for background, grounded on the inherent observation separation between foreground and background. However, a frequency spectrum entropy analysis suggests that this paradigm is not applicable to medical images where the foreground and background share numerous visual features, necessitating amore detailed description for the background. In this paper, we present anew Background-fused prototype (Bro) approach for FSS in medical images. Instead of identifying a commonality of background subjects in the support image, Bro fuses this background to discriminative prototypes, with two pivot designs. Specifically, Feature Similarity Calibration (FeaC) initially reduces noise in the support image by employing feature cross-attention with the query image. Subsequently, Hierarchical Channel-Adversarial Attention (HiCA) merges the background into comprehensive prototypes. We achieve this by a channel groups-based attention mechanism, where an adversarial Mean-Offset structure encourages a coarse-to-fine fusion. Designed as a generic plug-in, our Bro can be seamlessly integrated with existing FSS models. Extensive  \n∗ Corresponding author  \nEmail addresses: [tangs@usst.edu.cn](tangs@usst.edu.cn) (Song Tang), [zhaoyupumch@163.com](zhaoyupumch@163.com) (Yu  \nZhao)  \nexperiments validate the specificity of the background in medical images and the efficacy of Bro in enhancing the performance of previous FSS models on standard benchmarks.  \nKeywords: Medical image, Few-shot semantic segmentation, Adversarial regularization, Background-fused prototype, Channel group attention  \n1. Introduction  \nMedical image segmentation is a foundational task in clinical processes and medical research, with significant potential for various downstream applications such as disease diagnosis [1] and treatment planning [1] . Among the current topics, Few-shot Semantic Segmentation (FSS) [2] is an important area of focus, to account for the limited availability of well-annotated data, which arises from the protection of privacy and the requirement of clinical expertise. Unlike the conventional setting with segmentation labels, FSS’s objective is to predict the tissue or organ in query data, the same as the given one or several support data.  \nIn the view of building the similarity between the query and support images, the existing approaches mainly align to three lines: (1) The knowledge distillation framework [3](query and support images are inputs for the student and teacher branches, respectively), (2) the relevance structure discovering, e.g., attention [4] and graph [5], to identify shared features representing this similarity, and (3) the prototypical approach [6] to generate prototypes from support images to build this similarity with the query image. Since the prototypes capture the discriminative and robust visual factors while being compatible with the classic convolution computation pipeline, the prototypical paradigm is widely applied. In practical design, the previous prototypical methods engage in extracting the discriminative foreground prototype, the same as the scenarios of natural images, while background is represented by simplistic schemes, ","cbCaimC580n6oaYr","https://ap.wps.com/l/cbCaimC580n6oaYr","pdf",6141245,1,26,"English","en",105,"# Introduction\n## Few-shot semantic segmentation and limitations of foreground-focused prototypes\n## Motivation from frequency spectrum entropy\n# Proposed Method\n## Background-fused prototype (Bro)\n### Feature Similarity Calibration (FeaC)\n### Hierarchical Channel-Adversarial Attention (HiCA)\n# Experiments\n## Benchmark evaluation and results","[{\"question\":\"What problem does few-shot semantic segmentation aim to solve?\",\"answer\":\"It adapts a pretrained segmentation model to new classes using as few as a single labeled support sample per class.\"},{\"question\":\"Why are foreground-centered prototypical methods insufficient for medical images?\",\"answer\":\"Medical foreground and background share many visual features, so simplifying the background representation fails to capture discriminative differences.\"},{\"question\":\"How does the Bro approach improve medical image segmentation?\",\"answer\":\"Bro fuses background into discriminative prototypes using FeaC to reduce support noise via feature cross-attention and HiCA to merge background through channel group adversarial attention for coarse-to-fine fusion.\"}]",1784173864,66,{"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},"prototypical-few-shot-medical-image-semantic-segmentation-with-background-fusion","",{"@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/prototypical-few-shot-medical-image-semantic-segmentation-with-background-fusion/81506/",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","2026-07-25","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 few-shot semantic segmentation aim to solve?","Question",{"text":75,"@type":76},"It adapts a pretrained segmentation model to new classes using as few as a single labeled support sample per class.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Why are foreground-centered prototypical methods insufficient for medical images?",{"text":80,"@type":76},"Medical foreground and background share many visual features, so simplifying the background representation fails to capture discriminative differences.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the Bro approach improve medical image segmentation?",{"text":84,"@type":76},"Bro fuses background into discriminative prototypes using FeaC to reduce support noise via feature cross-attention and HiCA to merge background through channel group adversarial attention for coarse-to-fine fusion.","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"]