发表机构
Department of Biomedical Engineering, Baskent University; Department of Biomedical Engineering, Ankara University（巴塞特大学生物医学工程系; 安卡拉大学生物医学工程系）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究针对糖尿病视网膜病变早期病变检测难题，提出PRISM-DR管道，为每个病变训练单类检测器，经多步骤处理，在IDRiD数据集上训练后取得一定成果，证明将病变单独检测是实用替代方案。
AI 中文摘要
糖尿病视网膜病变是可预防失明的主要原因，其早期病变小、对比度低，易在人工筛查中被漏检。多数自动检测方法用单一多类模型处理四种非增殖性糖尿病视网膜病变，而这些病变在大小、颜色、形态和患病率上差异很大。我们提出PRISM-DR，一种针对病变的管道，为每个病变训练一个单类检测器，各有其配置。该管道从原始眼底图像开始，经过感兴趣区域裁剪、眼底特异性预处理、四个并行的YOLO检测器、平铺、五个交叉验证折叠的病变内集成，以及通过物理病变大小和临床优先级而非置信度来解决重叠的病变间抑制步骤。每个病变选择五个YOLO代中的最佳模型，并通过贝叶斯优化调整增强。在IDRiD上进行分层五折交叉验证训练后，系统在测试集上的mAP50为0.527，F1为0.529，在硬性渗出物上的AP50最高，为0.561。未经微调时，模型在成像比例接近IDRiD的情况下转移良好，随着视野和分辨率的偏离而退化。这些适度的绝对结果反映了单一来源训练集小且任务困难；然而，将每个病变视为单独的检测问题是单一多类模型的一种实用替代方案。
英文摘要
Diabetic retinopathy is a leading cause of preventable blindness; its early lesions are small, low contrast, and easily missed in manual screening. Most automated detectors handle the four non-proliferative DR lesions: microaneurysms, hemorrhages, hard exudates, and soft exudates, with a single multi-class model, even though these lesions differ sharply in size, color, morphology, and prevalence, so a shared model favors common, easy classes over rare, difficult ones. We present PRISM-DR, a lesion-specific pipeline that trains one single-class detector per lesion, each with its own configuration. From a raw fundus image, the pipeline applies region of interest cropping, fundus-specific preprocessing, four parallel YOLO detectors, tiling, per-lesion ensembling of five cross-validation folds, and an inter-lesion suppression step that resolves overlaps by physical lesion size and clinical priority rather than confidence. Per lesion, the best of five YOLO generations is selected, and augmentation is tuned by Bayesian optimization. Trained on IDRiD with stratified five-fold cross-validation, the system reaches a test mAP50 of 0.527 and F1 of 0.529, highest AP50 on hard exudates with 0.561. Without fine-tuning, the models transfer well where the imaging scale is close to IDRiD and degrade as field of view and resolution depart. These modest absolute results reflect a small single-source training set and a difficult task; however, treating each lesion as a separate detection problem is a practical alternative to a single multi-class model.
Comments17 pages, 9 figures, 15 tables
