RVD
收藏资源简介:
该数据集包含635个智能手机拍摄的眼底视频,来自四个不同诊所的415名50至75岁的患者。数据集提供了空间和时间维度上全面的视网膜结构精确标注,旨在推动血管分割领域的发展。具体包括三种空间标注:整体视网膜结构的二值血管掩码、区分静脉和动脉的通用静脉-动脉掩码以及进一步表征每个动脉和静脉粒度的精细静脉-动脉掩码。此外,数据集还提供了捕捉血管搏动特征的时间标注,有助于检测需要精细血流动力学波动识别的眼部疾病。
This dataset contains 635 fundus videos captured by smartphones, collected from 415 patients aged 50 to 75 years across four distinct clinics. It provides comprehensive and precise annotations of retinal structures across both spatial and temporal dimensions, aiming to advance the field of vascular segmentation. Specifically, it includes three types of spatial annotations: binary vascular masks for the overall retinal vasculature, general arteriovenous masks that distinguish between veins and arteries, and fine-grained arteriovenous masks that further characterize the granularity of each individual artery and vein. Additionally, the dataset offers temporal annotations that capture the characteristics of vascular pulsation, which facilitates the detection of ocular diseases requiring precise identification of hemodynamic fluctuations.

- 1RVD: A Handheld Device-Based Fundus Video Dataset for Retinal Vessel Segmentation · 2023年



