Fundus Image Myopia Development (FIMD) dataset

Description

Fundus Image Myopia Development (FIMD) dataset contains 70 retinal image pairs, in which, there is obvious myopia development between each pair of images. In addition, each pair of retinal images has a large overlap area, and there is no other retinopathy. In order to perform a reliable quantitative evaluation of registration results, we follow the annotation method of Fundus Image Registration (FIRE) dataset [1] to label control points between the pair of retinal images with the help of experienced ophthalmologists. Each image pair is labeled with 12 pairs of widely and uniformly distributed control points, which are mainly located at the bifurcation and crossing of blood vessels. FIMD also records the corresponding vision screening data of each image, including age, gender and uncorrected visual acuity. It should be noted that the time interval between two images is relatively long (one to five years), and the fundus cameras are different, which leads to different resolution, brightness, contrast and field of vision between two images. Moreover, the structure of our FIMD dataset is shown in "FIMD-dataset-structure.txt". You can conduct registration studies on image pairs in the dataset, and then evaluate the registration effect through annotated control points. In addition, we also record the age, gender, and uncorrected visual acuity for each image, so related study on the development of myopia may also be performed on this dataset. It depends on your own thinking. Reference: [1] Hernandez-Matas C, Zabulis X, Triantafyllou A, et al. FIRE: fundus image registration dataset[J]. Modeling and Artificial Intelligence in Ophthalmology, 2017, 1(4): 16-28.

Institutions

Categories

Funding

Licence