[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-83071-en":3,"doc-seo-83071-105":30,"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":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},83071,1374391975076,"Riley","https://ap-avatar.wpscdn.com/avatar/14000253ca4ec9f6853?x-image-process=image/resize,m_fixed,w_180,h_180&k=1783305029341752051",8,"Research & Report","MAC-XA Multi-view Anatomy-Correspondence Fusion for Coronary Stenosis Reporting from X-ray Angiography","Multi-view reasoning in coronary X-ray angiography is a cross-projection geometric task, but automated report generation is still limited. Projection-dependent branch overlap and foreshortening make single-view modeling incomplete and unstable for lesion localization and stenosis grading. Existing fusion methods learn correspondence implicitly because cross-view alignment is unobservable and cannot be explicitly supervised. This work reformulates reporting as alignment-constrained aggregation, adding controllable synthetic angiography for patch-level correspondence supervision and an anatomy-correspondence module for consistent cross-view feature alignment, improving structured stenosis reporting and zero-shot transfer to real angiograms.","arXiv :2607 .06268v 1 [ cs .CV] 7 Jul 2026  \nMAC-XA: Multi-view Anatomy-Correspondence Fusion for Coronary Stenosis Reporting from X-ray Angiography  \nChen Jia 1 ,6 , Baochang Zhang 1 ,2 ,3⋆, Fatia Kusuma Dewi 1 , Amir Yousefi 1 , Heribert Schunkert2 ,4 , Reza Ghotbi5 , and Nassir Navab 1 ,3  \n1 Computer Aided Medical Procedures, Technical University of Munich, Munich,  \nGermany  \n{chen.jia,[baochang.zhang}@tum.de](baochang.zhang}@tum.de)  \n2 German Heart Center Munich, Munich, Germany  \n3 Munich Center for Machine Learning, Munich, Germany  \n4 German Centre for Cardiovascular Research, Munich Heart Alliance, Munich, Germany  \n5 HELIOS Hospital West of Munich, Munich, Germany  \n6 relAI – Konrad Zuse School of Excellence in Reliable AI, Munich, Germany  \nAbstract. Multi-view reasoning in coronary X-ray angiography is inherently a cross-projection geometric problem, yet automated report generation in this setting remains largely unexplored. The 3D vascular topology leads to projection-dependent branch overlap and foreshortening, rendering single-view modeling fundamentally incomplete and unstable for lesion localization and stenosis grading. Although multi-view fusion appears promising, learning anatomically consistent fusion from real angiograms is impeded by a critical limitation: cross-view alignment is unobservable and cannot be explicitly supervised. Consequently, conventional fusion relies on implicit correlations rather than verified anatomical correspondence. We address this by reformulating multi-view stenosis reporting as an alignment-constrained aggregation problem. A controllable synthetic angiography generation strategy is introduced to expose geometry-derived patch-level correspondence supervision unavailable in real data. An anatomy-correspondence module learns cross-view correspondence matrices that explicitly align auxiliary features within the main-view coordinate space prior to fusion, thereby constraining evidence aggregation to anatomically consistent regions. Experiments on synthetic data and zero-shot transfer to real angiograms show that this alignment-constrained design improves correspondence consistency and structured stenosis reporting compared to single-view modeling and conventional multi-view fusion methods. The code will be publicly available upon publication.  \nKeywords: Medical Report Generation · Coronary Stenosis Reporting  \n· Anatomy-Correspondence Fusion · Multi-view X-ray Angiography.  \n⋆ Corresponding author.  \n2 C. Jia et al.  \n1 Introduction  \nCoronary artery stenosis is a major driver of coronary artery disease and requires precise localization and grading [1, 13] . Projection-based X-ray imaging compresses 3D anatomy into a 2D plane, removing depth cues and introducing view-dependent distortion [6, 15] . In coronary angiography, thin vessels, uneven contrast, and device interference further exacerbate projection-induced branch overlap and foreshortening, causing lesion visibility to vary substantially across views [3, 18] . Near vascular bifurcations, similar local appearance combined with frequent overlap makes single-view branch identification particularly unreliable [9, 5, 7] . Reliable stenosis reporting therefore requires integrating evidence across projections and reasoning over consistent vascular topology. Despite this clinical reality, automated multi-view report generation in coronary angiography remains largely underexplored.  \nRelated work has explored multi-view report generation and fusion mainly inchest X-ray settings, typically fusing images via attention or contrastive objectives and learning cross-view interactions implicitly from the end-task loss. For example, multi-view contrastive learning can strengthen representations and provide auxiliary guidance for transformer decoding [2, 10] . Other frameworks first derive prior concepts and then decode reports from fused multi-view knowledge [11] . More generally, multi-view fusion designs such as Duoduo CLIP [8] aggregat","cbCaijhBv8W80cRc","https://ap.wps.com/l/cbCaijhBv8W80cRc","pdf",4318797,3,1,10,"English","en",105,"# Introduction\n## Contributions\n# Method","[{\"question\":\"Why is single-view modeling unreliable for coronary stenosis reporting in X-ray angiography?\",\"answer\":\"Projection-based imaging compresses 3D anatomy into 2D, introducing view-dependent distortion and projection-induced branch overlap and foreshortening. As a result, lesion visibility and local appearance vary across views, especially near bifurcations where overlap is frequent.\"},{\"question\":\"What limitation prevents conventional multi-view fusion from enforcing anatomical correspondence?\",\"answer\":\"Cross-view alignment in real angiograms is unobservable, so it cannot be explicitly supervised. Conventional fusion therefore relies on implicit correlations rather than verified anatomical correspondence.\"},{\"question\":\"How does MAC-XA improve correspondence learning and reporting accuracy?\",\"answer\":\"MAC-XA reformulates stenosis reporting as an alignment-constrained aggregation problem and introduces controllable synthetic angiography to provide geometry-derived patch-level correspondence supervision. An anatomy-correspondence module predicts correspondence matrices to explicitly align auxiliary features in the main-view coordinate space before fusion, improving correspondence consistency and structured stenosis reporting, including zero-shot transfer to real angiograms.\"}]",1784185007,25,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"mac-xa-multi-view-anatomy-correspondence-fusion-for-coronary-stenosis-reporting-from-x-ray-angiography","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,50],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":20},"https://docshare.wps.com/document/research-report/",{"item":51,"name":13,"@type":43,"position":52},"https://docshare.wps.com/document/mac-xa-multi-view-anatomy-correspondence-fusion-for-coronary-stenosis-reporting-from-x-ray-angiography/83071/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-24","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},"Why is single-view modeling unreliable for coronary stenosis reporting in X-ray angiography?","Question",{"text":75,"@type":76},"Projection-based imaging compresses 3D anatomy into 2D, introducing view-dependent distortion and projection-induced branch overlap and foreshortening. As a result, lesion visibility and local appearance vary across views, especially near bifurcations where overlap is frequent.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What limitation prevents conventional multi-view fusion from enforcing anatomical correspondence?",{"text":80,"@type":76},"Cross-view alignment in real angiograms is unobservable, so it cannot be explicitly supervised. Conventional fusion therefore relies on implicit correlations rather than verified anatomical correspondence.",{"name":82,"@type":73,"acceptedAnswer":83},"How does MAC-XA improve correspondence learning and reporting accuracy?",{"text":84,"@type":76},"MAC-XA reformulates stenosis reporting as an alignment-constrained aggregation problem and introduces controllable synthetic angiography to provide geometry-derived patch-level correspondence supervision. An anatomy-correspondence module predicts correspondence matrices to explicitly align auxiliary features in the main-view coordinate space before fusion, improving correspondence consistency and structured stenosis reporting, including zero-shot transfer to real angiograms.","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":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,134],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":22,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":22,"slug":133},"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]