Datasets
These datasets are in the public domain
OCTID: Optical Coherence Tomography Image Database. This is a public database of over 500 OCT retinal images classified into 4 groups (including normals) as well as 25 manually segmented images and a semi-automated method for manual segmentation.
http://doi.org/10.3886/E108503V1
This is a de-identified dataset of retinal fundus images for glaucoma analysis (RIGA) derived from three sources for a total of 750 images. The optic cup and disc boundaries for each image are marked and were annotated manually by six experienced ophthalmologists and includes the cup to disc (CDR) estimates. Six parameters are extracted and assessed (the disc area and centroid, cup area and centroid, horizontal and vertical cup to disc ratios) among the ophthalmologists.
https://deepblue.lib.umich.edu/data/concern/data_sets/3b591905z
This is a completely artificially generated dataset of fundus images with their corresponding blood vessel annotation
https://www.openicpsr.org/openicpsr/project/117290/version/V1/view
The Foveal Avascular Zone (FAZ) is of clinical importance since the vascular arrangement around the fovea changes with disease and refractive state of the eye. In order to test and validate newly developed automated segmentation algorithms, we have created a public dataset of these retinal fundus images consisting of a total of 304 different images classified into: Diabetic (107), Myopic (109) and Normal (88) eyes. The images are of dimensions 420 x 420 pixels corresponding to 6mm x 6mm dimension of the retina. For each type of image, clear and manually segmented by a clinical expert (ground truth) are available.