ORVS (Online Retinal image for Vessel Segmentation (ORVS))

ORVS 数据集是卡尔加里大学计算机科学和视觉科学系之间的合作新建立的。
该数据集包含从加拿大卡尔加里的一家诊所收集的 49 幅图像（42 幅训练图像和 7 幅测试图像）。所有图像均使用具有 30 度视野 (FOV) 的 Zeiss Visucam 200 获取。图像大小为 1444×1444，每像素 24 位。图像以 JPEG 格式存储，压缩率低，这在眼科实践中很常见。所有图像均由一位一直在视网膜图像分析领域工作并经过培训的专家手动追踪。要求专家标记属于视网膜血管的所有像素。 Windows Paint 3D 工具用于手动标记图像。

The ORVS dataset was newly established through a collaborative partnership between the Department of Computer Science and the Department of Vision Sciences at the University of Calgary. This dataset comprises 49 images acquired from a clinic in Calgary, Canada, with 42 designated for training and 7 for testing. All images were captured using a Zeiss Visucam 200 with a 30-degree field of view (FOV). The images possess a spatial resolution of 1444 × 1444 and a bit depth of 24 bits per pixel. They are stored in JPEG format with low compression, a practice commonly adopted in ophthalmic clinical workflows. All images were manually annotated by an expert with extensive experience and specialized training in retinal image analysis, who was tasked with labeling all pixels corresponding to retinal blood vessels. The Windows Paint 3D tool was employed to complete the manual annotation process for these images.

搜集汇总

背景与挑战

背景概述

ORVS数据集是一个用于视网膜血管分割的医疗图像数据集，包含49幅高分辨率图像（42训练，7测试），由专家手动标记，适用于血管分割研究。数据集由卡尔加里大学等机构合作建立，2020年发布。