[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-153209-105":59,"doc-detail-153209-en":131},{"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":124,"head_meta":126,"extra_data":128,"updated_unix":130},105,"en","addressing-high-class-imbalance-in-multi-class-diabetic-retinopathy-severity-grading-with-augmentation-and-transfer-learning","Addressing High Class Imbalance in Multi-Class Diabetic Retinopathy Severity Grading with Augmentation and Transfer Learning","","Diabetic retinopathy (DR) is a leading cause of vision loss worldwide, and early diagnosis through automated retinal image analysis can reduce the risk of blindness. This paper presents a deep learning framework for binary and five-class DR severity classification using transfer learning and extensive class-balanced data augmentation. Experiments evaluate pretrained CNN variants, including ResNet and EfficientNet, on the APTOS 2019 dataset, improving performance under severe class imbalance and limited training data.",{"@graph":69,"@context":123},[70,84,106],{"@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/addressing-high-class-imbalance-in-multi-class-diabetic-retinopathy-severity-grading-with-augmentation-and-transfer-learning/153209/",{"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/addressing-high-class-imbalance-in-multi-class-diabetic-retinopathy-severity-grading-with-augmentation-and-transfer-learning/153209.png","ImageObject",300,407,{"name":92,"@type":93},"Theodore","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-27","2026-08-27",true,{"@type":102,"interactionType":103,"userInteractionCount":105},"InteractionCounter",{"@type":104},"ViewAction",14,{"@type":107,"mainEntity":108},"FAQPage",[109,115,119],{"name":110,"@type":111,"acceptedAnswer":112},"What problem does the paper address in diabetic retinopathy grading?","Question",{"text":113,"@type":114},"The paper focuses on high class imbalance in multi-class DR severity grading and how it affects model bias toward majority stages.","Answer",{"name":116,"@type":111,"acceptedAnswer":117},"What techniques are used to improve classification performance?",{"text":118,"@type":114},"It combines class-balanced data augmentation with transfer learning and fine-tuning of pretrained CNNs such as EfficientNet and ResNet.",{"name":120,"@type":111,"acceptedAnswer":121},"How well does the proposed framework perform for binary and five-class tasks?",{"text":122,"@type":114},"For binary classification it reaches 98.9% accuracy and 99.4% AUC. For five-class severity grading it achieves 84.6% accuracy and 94.1% AUC.","https://schema.org",{"og:url":83,"og:type":125,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":127,"canonical":83},"index,follow",{"doc_id":129,"site_id":62},153209,1787872064,{"code":4,"msg":5,"data":132},{"doc_id":129,"user_id":133,"nickname":92,"user_avatar":134,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":135,"file_id":136,"file_url":137,"file_type":138,"file_size":139,"view_count":105,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":52,"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":130,"read_time":144,"preview_page_count":4},7971461740886,"https://ap-avatar.wpscdn.com/davatar_3d24733baf745e90a7e4bdd5f77d97b2","arXiv :2507 . 17121v2 [ cs .CV] 28 Jul 2025  \nHighlights  \nAddressing High Class Imbalance in Multi-Class Diabetic Retinopathy Severity Grading with Augmentation and Transfer Learning  \nFaisal Ahmed  \n• A deep learning framework is developed for diabetic retinopathy (DR) severity classification, addressing both binary and five-class tasks.  \n• A class-balanced data augmentation strategy is introduced, generating 20,000 synthetic samples per class to mitigate data imbalance.  \n• Transfer learning and fine-tuning of pretrained CNNs (EfficientNet, ResNet) are extensively evaluated to identify optimal architectures for DR classification.  \n• The proposed model achieves state-of-the-art binary classification performance with 98.9% accuracy and 99.4% AUC.  \n• Competitive results are obtained for five-class DR grading, reaching 84.6% accuracy and 94.1%  \nAUC.  \nAddressing High Class Imbalance in Multi-Class Diabetic Retinopathy Severity Grading with Augmentation and Transfer Learning  \nFaisal Ahmeda,∗  \na Department of Data Science and Mathematics, Embry-Riddle Aeronautical University, 3700 Willow Creek Rd, Prescott, Arizona 86301, USA  \nAbstract  \nDiabetic retinopathy (DR) is a leading cause of vision loss worldwide, and early diagnosis through automated retinal image analysis can significantly reduce the risk of blindness. This paper presents a robust deep learning framework for both binary and five-class DR classification, leveraging transfer learning and extensive data augmentation to address the challenges of class imbalance and limited training data. We evaluate a range of pretrained convolutional neural network architectures, including variants of ResNet and EfficientNet, on the APTOS 2019 dataset.  \nFor binary classification, our proposed model achieves a state-of-the-art accuracy of 98.9%, with a precision of 98.6%, recall of 99.3%, F1-score of 98.9%, and an AUC of 99.4% . In the more challenging five-class severity classification task, our model obtains a competitive accuracy of 84.6% and an AUC of 94.1%, outperforming several existing approaches. Our findings also demonstrate that EfficientNet-B0 and ResNet34 offer optimal trade-offs between accuracy and computational efficiency across both tasks.  \nThese results underscore the effectiveness of combining class-balanced augmentation with transfer learning for high-performance DR diagnosis. The proposed framework provides a scalable and accurate solution for DR screening, with potential for deployment in real-world clinical environments.  \nKeywords: Diabetic Retinopathy, Deep Learning, Transfer Learning, Data Augmentation,, Medical Imaging  \n1. Introduction  \nAs of August 2023, the World Health Organization (WHO) reports that more than 2.2 billion people globally suffer from near or distance vision impairment, with at least 1 billion of these cases being preventable or unaddressed [1] . Among the leading causes of vision loss is diabetic retinopathy (DR), a common microvascular complication of diabetes that damages the blood vessels in the retina. DR affects an estimated 3.9 million people worldwide and, if left undetected, can progress from mild retinal abnormalities to complete blindness. Early diagnosis and grading of DR severity are crucial for initiating timely treatment and preventing irreversible vision damage.  \nTraditionally, DR diagnosis is performed through manual inspection of color fundus images by trained ophthalmologists. This process is labor-intensive, timeconsuming, and subject to inter-observer variability,  \n∗ Corresponding author  \nEmail address: [ahmedf9@erau.edu](ahmedf9@erau.edu) (Faisal Ahmed)  \nwhich may lead to inconsistent grading results and delayed treatment. In response, the medical imaging community has increasingly adopted machine learning (ML) techniques—particularly deep learning (DL)—to automate and improve DR detection and severity classification [2, 3] . Convolutional neural networks (CNNs), in particular, have shown impressive performan","cbCaitBd9iKqOl7i","https://ap.wps.com/l/cbCaitBd9iKqOl7i","pdf",228979,"English","# Introduction\n## Problem context: diabetic retinopathy and vision loss\n## Motivation: class imbalance in public datasets\n## Proposed approach: augmentation and transfer learning\n## Evaluation setup: APTOS 2019 and CNN backbones\n## Results for binary and five-class grading","[{\"question\":\"What problem does the paper address in diabetic retinopathy grading?\",\"answer\":\"The paper focuses on high class imbalance in multi-class DR severity grading and how it affects model bias toward majority stages.\"},{\"question\":\"What techniques are used to improve classification performance?\",\"answer\":\"It combines class-balanced data augmentation with transfer learning and fine-tuning of pretrained CNNs such as EfficientNet and ResNet.\"},{\"question\":\"How well does the proposed framework perform for binary and five-class tasks?\",\"answer\":\"For binary classification it reaches 98.9% accuracy and 99.4% AUC. For five-class severity grading it achieves 84.6% accuracy and 94.1% AUC.\"}]","Addressing High Class Imbalance in Multi-Class Diabetic Retinopathy Severity Grading with Augmentation and Transfer Learning | PDF",25]