[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-83255-en":3,"doc-seo-83255-105":29,"detail-sidebar-cat-0-en-105":90},{"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":4,"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},83255,13056703019662,"Evangeline","https://ap-avatar.wpscdn.com/avatar/be000253a8e92610077?_k=1778726343310543188",7,"Healthcare","Learning to Unify Deformable Shape and Texture Representations for Cardiac Video Classification","Deformable shape representations complement texture features in cardiac image classification by providing geometric priors robust to imaging artifacts and intensity variations. Existing deep networks often concatenate features without leveraging cross-modal dependencies, and they apply uniform attention over all timepoints, overlooking differences in diagnostic relevance across cardiac phases. ShapeFuse introduces a bidirectional cross-attention fusion in a shared latent space, enabling temporal, spatio-temporal correspondence–aware weighting of shape and texture over time. Experiments on cine CMR videos achieve state-of-the-art performance and improved interpretability via attention.","arXiv :2607 .075 18v 1 [ cs .CV] 8 Jul 2026  \nLearning to Unify Deformable Shape and Texture Representations for Cardiac Video Classification  \nTonmoy Hossain 1 and Miaomiao Zhang 1 ,2  \n1 Department of Computer Science, University of Virginia, Virginia, USA  \n2 Department of Electrical and Computer Engineering, University of Virginia, USA  \nAbstract. Deformable shape representations have proven to be robust complements to texture features in cardiac image classification, offering geometric priors that are invariant to imaging artifacts and intensity variations. However, existing deep networks perform simple concatenation to combine these distinct feature representations, which neither fully exploits their complementary nature nor learns cross-modal feature dependencies. Furthermore, this results in uniform attention across all timepoints; hence ignoring the varying diagnostic importance across the cardiac phases. In this paper, we propose a novel cardiac video classification model that, for the first time, learns temporal features in an integrated space of deformable shape and image texture representations. In particular, we design a bi-directional cross-attention in the latent space to fuse latent deformable shape and image features, allowing each modality to adaptively weight the other based on spatio-temporal correspondence.  \nIn contrast to current methods that apply uniform weighting across all the cardiac phases, our approach learns to dynamically adjust the contributions of shape and texture representations, derived from images, over time. We demonstrate state-of-the-art classification performance on a cine cardiac magnetic resonance (CMR) video dataset, achieving improved interpretability from attention mechanisms that identify diagnostically critical cardiac phases and modality contributions. Our code is publicly available at [https://github.com/tonmoy-hossain/ShapeFuse](https://github.com/tonmoy-hossain/ShapeFuse).  \nKeywords: CMR Video Classification · Feature Fusion · Deformations.  \n1 Introduction  \nCardiac MRI enables non-invasive assessment of myocardial function and is widely used for disease diagnosis, yet automated classification from CMR videos remains challenging [8,18,19,29] . Recent models capture temporal dependencies across frames but solely operate on raw image intensities. As a result, they mostly focus on texture-based features without understanding the underlying myocardial deformations [2,10,32] . However, pathological conditions, such as cardiomyopathy and myocarditis manifest as regional wall motion abnormalities, asymmetric contractility, and altered strain patterns [12,13] . Therefore, geometric shape features captured in motion or deformation fields often encode richer information that can complement intensity-based representations alone [22,30,31] .  \n2 Hossain and Zhang  \nRecent works have explored integrating deformable shape representations with intensity features for neurodegenerative disease classification, where geometric priors have proven to be robust complements to texture-based representations [15,16,25] . Extending this paradigm to cardiac video analysis, however, is not straightforward given the temporal nature of the cardiac cycle. A critical challenge lies in how to combine these two intertwined representations, as simple operations of element-wise addition or concatenation can ignore cross-modal feature dependencies. In particular, such approaches do not model how shape and texture jointly interact across the cardiac cycle, which limits their ability to learn complementary physiological information.  \nFor instance, deformation features are most informative during dynamic transition phases of the cardiac cycle, particularly from end-diastole through peak systolic contraction and early relaxation, where myocardial motion changes are most evident. In contrast, texture-based intensity patterns are less sensitive to these functional dynamics and may become unreliable under conditio","cbCailk9En5Rq68b","https://ap.wps.com/l/cbCailk9En5Rq68b","pdf",1412674,1,11,"English","en",105,"# Introduction\n## Deformation learning from image pairs\n## ShapeFuse: unified latent fusion framework","[{\"question\":\"What limitation do current cardiac video models have when fusing deformable shape and texture features?\",\"answer\":\"They typically use simple concatenation and learn uniform attention across all timepoints, which fails to exploit cross-modal dependencies and ignores how diagnostic importance varies across cardiac phases.\"},{\"question\":\"How does ShapeFuse combine deformable shape and image texture representations?\",\"answer\":\"It learns temporal features in an integrated latent space using a bidirectional cross-attention module, where each modality adaptively reweights the other based on spatio-temporal correspondence.\"},{\"question\":\"What benefits do the learned attention mechanisms provide in ShapeFuse?\",\"answer\":\"They improve interpretability by highlighting diagnostically critical cardiac phases and quantifying modality-wise contributions to the final classification.\"}]",1784186298,28,{"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":85,"head_meta":87,"extra_data":89,"updated_unix":27},"learning-to-unify-deformable-shape-and-texture-representations-for-cardiac-video-classification","",{"@graph":35,"@context":84},[36,53,67],{"@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/healthcare/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/learning-to-unify-deformable-shape-and-texture-representations-for-cardiac-video-classification/83255/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"What limitation do current cardiac video models have when fusing deformable shape and texture features?","Question",{"text":74,"@type":75},"They typically use simple concatenation and learn uniform attention across all timepoints, which fails to exploit cross-modal dependencies and ignores how diagnostic importance varies across cardiac phases.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"How does ShapeFuse combine deformable shape and image texture representations?",{"text":79,"@type":75},"It learns temporal features in an integrated latent space using a bidirectional cross-attention module, where each modality adaptively reweights the other based on spatio-temporal correspondence.",{"name":81,"@type":72,"acceptedAnswer":82},"What benefits do the learned attention mechanisms provide in ShapeFuse?",{"text":83,"@type":75},"They improve interpretability by highlighting diagnostically critical cardiac phases and quantifying modality-wise contributions to the final classification.","https://schema.org",{"og:url":51,"og:type":86,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":88,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":91},[92,96,100,104,109,114,117,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":45,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":115,"slug":116},40,"healthcare",{"id":118,"doc_module":4,"doc_module_name":45,"category_name":119,"show_sort_weight":120,"slug":121},8,"Research & Report",30,"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":105,"slug":137},19,"General","general"]