Dataset Description
Drishti-GS1 dataset consists of a total of 101 images.These have been divided into 50 training and 51 testing images.All the images have been marked by 4 eye experts with varying clinical experience. All images were collected at Aravind eye hospital, Madurai from visitors to the hospital, with their consent. Glaucoma patient selection was done by clinical investigators based on clinical findings during the visit. Selected patients were between 40-80 years of age with roughly equal number of males and females. The data collection protocol was as follows.
- All images were taken centered on OD with a Field-Of-View(FOV) of 30-degrees and of dimensions 2896 X 1944 pixels and PNG uncomressed image format.
- Groundtruth was collected from data experts with varying clinical experience of 3,5,9 and 20 years respectively.
- For the groundtruth collection a dedicated marking tool was created to allow for precise boundary marking.
- Segmentation Soft Map : It has been obtained by fusing the markings of all the 4 medical experts.
- Average OD and Cup boundaries : These are obtained by averaging the manual markings over many angular sectors.
- CDR Values : Since CDR is an important parameter for glaucoma assessment , hence CDR values corresponding to 4 expert markings are also provided.
- Image-level Decisions: A binary image-level decision on whether each image is normal or glaucomatous is provided based on the majority opinion (3 out of 4) of experts
- Notching : Decision on the occurence of notching in the superior, inferior, nasal, and temporal sectors, assessed by a single expert is provided
We provide the groundtruth for the 50 training images. For the testing images, individual teams can report their results after registering on this website and we will evaluate them and put up on this web site.
For a detailed description of the dataset you may refer to the paper.