• The fundus images in IDRiD were captured by a retinal specialist at an Eye Clinic located in Nanded, Maharashtra, India. From the thousands of examinations available, we have extracted 516 images to form our dataset.
  • Experts verified that all images are of adequate quality, clinically relevant, that no image is duplicated and that a reasonable mixture of disease stratification representative of diabetic retinopathy (DR) and diabetic macular edema (DME) is present.
  Fundus Camera Specifications

Images were acquired using a Kowa VX-10 alpha digital fundus camera with 50-degree field of view (FOV), and all are centered near to the macula. The images have a resolution of 4288×2848 pixels and are stored in jpg file format. The size of each image is about 800 KB. 

  Description of Data
The dataset provides expert markups of typical diabetic retinopathy lesions and normal retinal structures. Furthermore, it provides information regarding disease severity level of diabetic retinopathy, and diabetic macular edema for each image in the database based on international standards of clinical relevance. The details about available ground-truth are in following sections.
  Pixel Level Annotated Data
This dataset consists of 81 color fundus images with signs of DR. Precise pixel level annotation of abnormalities associated with DR like microaneurysms (MA), soft exudates (SE), hard exudates (EX) and hemorrhages (HE) is provided as a binary mask for performance evaluation of individual lesion segmentation techniques. It includes color fundus images (.jpg files) and binary masks made of lesions (.tif files). Number of images (some images contain multiple lesions) with binary masks available for particular lesion is given as follows: 
  • MA – 81
  • EX –  81
  • HE –  80
  • SE –  40
In addition to all the abnormalities, binary masks for the optic disc region are provided for all 81 images.
  Image Level Grading
The medical experts graded the full set of 516 images with a variety of pathological conditions of DR and DME. Grading for all images is available in the CSV file. The diabetic retinal images were classified into separate groups according to the International Clinical Diabetic Retinopathy Scale. The macular edema severity was decided based on occurrences of hard exudates near to macula center region.
Along with the annotations presented above, optic disc and fovea center locations are provided as separate CSV file.
  Data Download Link
Complete Data can be downloaded from IEEE Dataport Repository: Indian Diabetic Retinopathy Image Dataset (IDRiD)

This work is licensed under a Creative Commons Attribution 4.0 International License.