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.