首个开放的视网膜旁凹OCTA图像数据集

本数据集由爱丁堡大学休斯研究所创建，包含55个视网膜旁凹OCTA图像及其相应的地面实况手动分割。数据集旨在支持OCTA图像自动分割的研究，特别是在深度血管表型分析方面。数据集内容包括来自左右眼的图像，涵盖不同的临床兴趣区域，如上、鼻、下、颞和黄斑区域。创建过程中，研究人员采用了保守的像素级手动分割方法，以确保分割的准确性和可重复性。该数据集的应用领域主要集中在早期识别病理条件，如糖尿病视网膜病变、心血管疾病和神经退行性疾病，通过量化视网膜血管的结构和功能变化来辅助诊断和监测这些疾病。

This dataset was developed by the Hughes Institute at the University of Edinburgh, comprising 55 parafoveal retinal OCTA images and their corresponding manually generated ground truth segmentations. It is designed to support research on automated OCTA image segmentation, with a specific focus on deep vascular phenotyping analysis. The dataset includes images from both left and right eyes, covering a range of clinically relevant regions of interest: superior, nasal, inferior, temporal, and macular areas. During dataset construction, researchers adopted a conservative pixel-level manual segmentation workflow to ensure the accuracy and reproducibility of the segmentations. The primary application scenarios of this dataset center on the early identification of pathological conditions including diabetic retinopathy, cardiovascular diseases, and neurodegenerative disorders, assisting in the diagnosis and monitoring of these diseases by quantifying structural and functional changes in retinal blood vessels.

搜集汇总

背景与挑战

背景概述

该数据集是首个开放的视网膜旁凹光学相干断层扫描血管成像（OCTA）图像数据集，包含11名参与者的OCTA扫描图像及手动分割结果，图像采集自RTVue-XR Avanti系统，聚焦于浅层血管结构，旨在支持神经退行性疾病早期生物标志物的研究。