FIRE：眼底图像配准数据集

FIRE 是一个视网膜眼底图像数据集，包含 129张 眼底视网膜图像，由不同特征组合成 134对 图像组合。这些图像组合根据特质被划分为3类。眼底图像由 Nidek AFC-210 眼底照相机采集，分辨率为2912x2912，视觉仰角为40度。图像由 Papageorgiou Hospital 医院和Aristotle University of Thessaloniki大学共同构建，由于Thessaloniki 大学采集自39名患者。. 数据包括以下几部分内容： 如果使用数据集的任何部分，请引用以下文章： FIRE：眼底图像配准数据集 C. Hernandez-Matas，X。Zabulis，A。Triantafyllou，P。Anyfanti，S。Douma，AA Argyros Journal for Ophthalmology，vol。1，不。4，pp.16-28，Jul。2017。 该数据集由129个视网膜图像组成，形成134个图像对。这些图像对根据其特征分为3个不同的类别。使用Nidek AFC-210眼底照相机采集图像，该照相机在x和y维度上获取分辨率为2912×2912像素且FOV为45°的图像。图像来自39名患者在塞萨洛尼基塞萨洛尼基亚里士多德大学Papageorgiou医院。 图像遵循命名约定： [图像对名称] _X.jpg 其中参考图像的X为1，测试图像的X为2。 基础事实文件遵循命名约定： control_points_ [图像对名称] _1_2.txt 每个图像对的基础事实文件具有以下格式： [reference_point_1_x] [reference_point_1_y] [test_point_1_x] [test_point_1_y] [reference_point_2_x] [reference_point_2_y] [test_point_2_x] [test_point_2_y]

FIRE is a retinal fundus image dataset consisting of 129 fundus retinal images, which form 134 image pairs with varying features. These image pairs are categorized into 3 distinct classes based on their characteristics. The images were captured using a Nidek AFC-210 fundus camera, with a resolution of 2912×2912 pixels and a visual field angle of 40 degrees. The dataset was jointly developed by Papageorgiou Hospital and Aristotle University of Thessaloniki, and the images were collected from 39 patients by Aristotle University of Thessaloniki. If any portion of this dataset is used, please cite the following paper: FIRE: A Retinal Fundus Image Registration Dataset. C. Hernandez-Matas, X. Zabulis, A. Triantafyllou, P. Anyfanti, S. Douma, A. A. Argyros. *Journal for Ophthalmology*, vol. 1, no. 4, pp. 16-28, Jul. 2017.

This dataset comprises 129 retinal images that form 134 image pairs, which are divided into 3 distinct categories based on their features. The images were acquired using a Nidek AFC-210 fundus camera, which captures images with a resolution of 2912×2912 pixels in both x and y dimensions and a field of view (FOV) of 45°. All images were collected from 39 patients at Papageorgiou Hospital, Aristotle University of Thessaloniki, Thessaloniki.

Images follow the naming convention: [image_pair_name]_X.jpg, where X=1 refers to the reference image and X=2 refers to the test image.

Ground truth files follow the naming convention: control_points_[image_pair_name]_1_2.txt. The ground truth file for each image pair follows the format below: [reference_point_1_x] [reference_point_1_y] [test_point_1_x] [test_point_1_y] [reference_point_2_x] [reference_point_2_y] [test_point_2_x] [test_point_2_y]

搜集汇总

背景与挑战

背景概述

FIRE是一个专为眼底图像配准研究设计的医学数据集，包含129张高分辨率视网膜图像形成的134对图像组合，分为3个类别。这些图像由专业设备采集自39名患者，附带彩色ROI掩模、特征ROI掩模和标注点，适用于医学图像处理和分析研究。