Retinal Image Registration

Retinal image registration plays a pivotal role in disease diagnosis, image-guided surgery, monitoring of disease progression, and image fusion. The primary objective of image registration is to spatially align a moving (source) image with a fixed (target) image by determining a geometric transformation that accurately matches the corresponding features or structures between the two images. In the realm of ophthalmology, retinal image registration has become an indispensable tool, facilitating precise tracking of temporal changes in retinal anatomy and the alignment of different imaging modalities, which is crucial for evaluating disease evolution and treatment outcomes.

We propose a new dataset for retinal image registration named Coph100 here. You can download the complete dataset from Baidu netdisk and google drive. Baidu Netdisk: COph100_V1 链接: https://pan.baidu.com/s/101xMvgKw-qDFCYiTnXZzvw 提取码: 5yv2 google drive: https://drive.google.com/drive/folders/1soVkFxUhbJBycUszR1nw6hky6_LTgZjH?usp=sharing

We will public the dataset after the paper is accept: figsharehttps://doi.org/10.6084/m9.figshare.27061084

Our dataset is constructed based on the ROP classification dataset proposed by Timkovic et al. (Timkovič, Juraj, et al. “Retinal Image Dataset of Infants and Retinopathy of Prematurity.” Scientific Data 11.1 (2024): 814.) Following the readme, you can obtain the registration dataset from figshare (need to download ROP classification dataset and ours) or you can download directly from baidu netdisk or google drive (with fixed and moving images and lables).

Please cite the paper: Hu, Y., Gong, M., Qiu, Z. et al. COph100: A comprehensive fundus image registration dataset from infants constituting the “RIDIRP” database. Sci Data 12, 99 (2025). https://doi.org/10.1038/s41597-025-04426-w