NETRA Dataset

The NETRA dataset is a comprehensive and detailed resource, specifically curated to facilitate the development and evaluation of automated algorithms for glaucoma diagnosis. This dataset originates from the B.P. Koirala Lions Centre for Ophthalmic Studies (BPKLOCS) Hospital in Nepal, a leading institution in the field of ophthalmology. The dataset consists of 519 high-resolution retinal images (3456 x 2304 pixels). Each of these images has been manually annotated by an experienced ophthalmologist. The annotation process, which was meticulously carried out using a tablet and pen, includes various pathological signs such as cup notching, rim thinning, and a fully cupped optic disc. This attention to detail underscores the quality and reliability of the dataset. The NETRA dataset covers an extensive variety of cases, ranging from mild to advanced glaucoma. This wide range enhances the value of the dataset, making it ideal for creating robust and versatile algorithms capable of identifying a broad spectrum of glaucoma conditions. The primary goal of the collection and distribution of the NETRA dataset is to promote the development of next-generation tools for mass glaucoma screening. It is designed to be used for training and evaluating algorithms that can accurately diagnose glaucoma, thereby significantly contributing to advancements in ophthalmology and healthcare.NETRA数据集是一套全面详尽的专业资源，专为助力青光眼（glaucoma）诊断自动化算法的开发与评估而精心打造。该数据集源自尼泊尔眼科领域顶尖机构——B.P.柯伊拉腊狮子眼科研究中心（B.P. Koirala Lions Centre for Ophthalmic Studies, BPKLOCS）医院。数据集包含519张高分辨率视网膜图像，分辨率为3456×2304像素，所有图像均由经验丰富的眼科医生完成手动标注。标注过程采用数位板与手写笔细致严谨地开展，涵盖杯切迹、视盘边缘变薄以及完全凹陷的视盘（optic disc）等多种病理特征。这种对细节的高度重视，凸显了该数据集的质量与可靠性。NETRA数据集覆盖了从轻度到重度青光眼的各类病例，丰富的样本覆盖范围极大提升了数据集的应用价值，使其非常适合用于开发可识别多种青光眼病症的稳健且通用的算法。本数据集的收集与发布，核心目标是推动下一代青光眼大规模筛查工具的研发，其设计初衷是用于训练和评估能够精准诊断青光眼的算法，从而为眼科医学与医疗健康领域的发展进步作出重要贡献。