[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-352356-105":59,"doc-detail-352356-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","hybrid-vision-transformer-and-graph-neural-network-model-with-region-adaptive-attention-for-enhanced-skin-cancer-prediction","Hybrid vision transformer and graph neural network model with region-adaptive attention for enhanced skin cancer prediction","","A well-known, potentially lethal skin cancer demands prompt detection and accurate diagnosis, yet conventional deep models struggle with complex lesion spatial linkages and global context in dermoscopy images. A Hybrid Vision Transformer (ViT) combined with a Graph Neural Network (GNN) and Region-Adaptive Attention is proposed for multi-scale lesion analysis, capturing global dependencies, spatial relationships, and diagnostically salient regions. Meta-learning supports generalization across skin tones and imaging settings.",{"@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/hybrid-vision-transformer-and-graph-neural-network-model-with-region-adaptive-attention-for-enhanced-skin-cancer-prediction/352356/",{"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/hybrid-vision-transformer-and-graph-neural-network-model-with-region-adaptive-attention-for-enhanced-skin-cancer-prediction/352356.png","ImageObject",300,407,{"name":92,"@type":93},"Noah","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-11-03","2026-09-22",true,{"@type":102,"interactionType":103,"userInteractionCount":52},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"Why are CNNs challenging for skin lesion diagnosis in this work?","Question",{"text":112,"@type":113},"CNNs can miss long-range correlations, have limited receptive fields, and struggle with irregular lesion architecture and complex spatial dependencies in dermoscopy images.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"How does the model combine ViT, GNN, and Region-Adaptive Attention?",{"text":117,"@type":113},"The ViT branch captures global contextual dependencies, the GNN encodes spatial and topological relations via a derived graph, and Region-Adaptive Attention improves feature extraction by focusing on diagnostically salient locations.",{"name":119,"@type":110,"acceptedAnswer":120},"How is the model evaluated for reliability and generalization?",{"text":121,"@type":113},"Performance is verified on two disparate real-world datasets, HAM10000 and ISIC, with cross-dataset testing supporting generalization beyond a single cohort.","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},352356,1790385236,{"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":52,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":139,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":144,"read_time":145,"preview_page_count":19,"preview_urls":146},8796095462418,"https://ap-avatar.wpscdn.com/avatar/80000253c1241d02b47?x-image-process=image/resize,m_fixed,w_180,h_180&k=1778826106357471780","[www. nature.com/scientificreports](www. nature.com/scientificreports)  \nOPEN  \nHybrid vision transformer and graph neural network model with region-adaptive attention for enhanced skin cancer prediction  \nAswani Dogga1,4􀀍, Sivasubramanian R.2,4 & Shanthi S.3,4  \nA well-known and potentially lethal skin cancer requires prompt detection and diagnosis. Complex spatial linkages and global contextual information in skin lesion photos challenge CNNs and other deep learning methods. Given these restrictions, we present a Hybrid Vision Transformer (ViT) with a Graph Neural Network (GNN) and Region-Adaptive Attention to diagnose skin cancer. The ViT branch captures dermoscopy image global dependencies, whereas the GNN enhances features by exploiting lesions’ spatial relationships. Region-Adaptive Attention improves lesion categorization by dynamically improving feature extraction in diagnostically relevant locations. Our paradigm for multi-scale lesion analysis accounts for lesion size, color, and texture changes. Meta-learning methods refine the proposed model to make it generalizable across skin tones and imaging settings. Our model outperformed state-of-the-art deep learning algorithms on benchmark skin cancer datasets. The architecture improves classification accuracy and interpretability, making it a promising clinical dermatology tool.  \nKeywords Skin cancer prediction, Vision transformer (ViT), Graph neural network (GNN), Regionadaptive attention (RAA), Deep learning, Medical image analysis, Multi-Scale lesion analysis  \nSkin cancer is considered to be among the most fatal cancers, with melanoma being categorized as the most deadly and aggressive type. Timely intervention considerably improves survival rates; hence, early detection and accurate diagnosis are important for improving treatment outcome1. Clinical dermatology relies on visual examination of skin lesions as its mainstay of diagnosis, yet this method is subjective and prone to errors. In a bid to overcome some of these limitations, the automatic detection and diagnosis for skin cancer using deep learning-based computer-aided diagnosis techniques had gained much research attention. Recently, convolutional neural networks (CNN) have proved to be very effective in the field of medical image diagnosis. Nonetheless, CNNs suffer some challenges including small receptive fields, long-range correlation capture, and irregular lesion architecture. Option alternatives are the Vision Transformers (ViT), which with better global contextual understanding than ConvNets, achieve superior results in medical imaging2. Unfortunately, ViTslack the ability to model spatial interactions among sub regions, which leads to their inability to understand the complex shapes of skin lesions. We thus propose a skin cancer prediction model that implements a Hybrid Vision Transformer (ViT) with a Graph Neural Network (GNN) incorporating Region-Adaptive Attention (RAA) . The GNN attempts to enhance the representation of atypical skin lesions by encoding spatial interactions between lesion territories while the ViT exploits global dependency mapping on dermoscopic images. Our RegionAdaptive Attention (RAA) thus increases the classification score by improving feature extraction with focus of attention on diagnostically salient areas3. Unlike classic approaches that rely solely on stacked convolutional networks or transformer-only pipelines, the proposed architecture leverages the complementary strengths of Vision Transformers and graph neural networks in a tightly woven, ViT-GNN hybrid design4. By deriving a graph from the Vision Transformer’s output maps, the framework explicitly learns the spatial and topological relations among lesion components, addressing the overlooked inter-region dependencies that remain absent inmost state-of-the-art pipelines. To further sharpen the model’s focus on clinically salient lesion characteristics, a custom hybrid attention mechanism is employed, while complementar","cbCaibMr5P7I5jrG","https://ap.wps.com/l/cbCaibMr5P7I5jrG","pdf",3774451,22,"English","# Introduction\n## Motivation and challenges in skin cancer diagnosis\n## Proposed Hybrid ViT-GNN framework\n# Model components\n## Embeddings-to-graph transformation\n## Region-Adaptive Attention and attention fusion\n# Performance and validation\n## Dataset evaluation and generalization\n## Interpretability modules","[{\"question\":\"Why are CNNs challenging for skin lesion diagnosis in this work?\",\"answer\":\"CNNs can miss long-range correlations, have limited receptive fields, and struggle with irregular lesion architecture and complex spatial dependencies in dermoscopy images.\"},{\"question\":\"How does the model combine ViT, GNN, and Region-Adaptive Attention?\",\"answer\":\"The ViT branch captures global contextual dependencies, the GNN encodes spatial and topological relations via a derived graph, and Region-Adaptive Attention improves feature extraction by focusing on diagnostically salient locations.\"},{\"question\":\"How is the model evaluated for reliability and generalization?\",\"answer\":\"Performance is verified on two disparate real-world datasets, HAM10000 and ISIC, with cross-dataset testing supporting generalization beyond a single cohort.\"}]","Hybrid vision transformer and graph neural network model with region-adaptive attention for enhanced skin cancer prediction | PDF",1790099159,55,[147,151,155,159],{"page_no":8,"url":148,"width":149,"height":150},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/7923489b7302938b7b3531c8fb28ba52.webp",983,1291,{"page_no":14,"url":152,"width":153,"height":154},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/9c5f09c28100ed0f37d2e9fb6eeabefe.webp",982,1290,{"page_no":81,"url":156,"width":157,"height":158},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/355a6ef769279ae074152e930815a380.webp",981,1289,{"page_no":19,"url":160,"width":161,"height":162},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/f0760d9bb4b8f43237838c8f927dd551.webp",980,1287]