COph100
收藏资源简介:
COph100是由南方科技大学等机构创建的婴儿视网膜图像配准数据集,包含100只眼睛的491对图像,数据来源于早产儿视网膜病变(ROP)患者。该数据集通过手动标注真实图像点和自动血管分割掩码,提供了丰富的图像质量问题和临床挑战。数据集构建过程中,研究人员从公开数据集中精心挑选了图像,并考虑了图像分辨率、光照等多样性因素。COph100的应用领域主要集中在儿科眼科疾病的诊断和治疗,特别是通过图像配准技术跟踪疾病进展,提升临床应用的鲁棒性和准确性。
COph100 is an infant retinal image registration dataset developed by Southern University of Science and Technology and other institutions. It consists of 491 pairs of images from 100 eyes, sourced from patients with retinopathy of prematurity (ROP). Equipped with manually annotated ground-truth image points and automatically generated vascular segmentation masks, this dataset covers a wide range of image quality issues and clinical challenges. During its construction, researchers carefully selected images from public datasets, taking into account diversity factors such as image resolution and illumination. The main application fields of COph100 focus on the diagnosis and treatment of pediatric ophthalmic diseases, particularly tracking disease progression via image registration techniques to enhance the robustness and accuracy of clinical applications.

- 1COph100: A comprehensive fundus image registration dataset from infants constituting the "RIDIRP" database南方科技大学可信自主系统研究院与计算机科学与工程系, 北京同仁医院眼科中心, 中国科学院深圳先进技术研究院高性能计算中心 · 2025年



