An AS-OCT image dataset for deep learning-enabled segmentation and 3D reconstruction for keratitis

This dataset contains a total of 1168 AS-OCT images of patients with keratitis, including 768 full-frame images (6 patients). Each image has associated segmentation labels for lesions and cornea, and also iris for full-frame images. The 2 categories of images are separated into two folders, named “Partial-frame_Dataset” and “Full-frame_Dataset” respectively. In each folder, original AS-OCT images in the BMP format are provided in the “Original_AS-OCT_Images” folder named as “n.bmp”, and corresponding annotated files in the JSON format are provided in the “Experts_Annotations” folder named as “n.json”. Anonymized information about participants was also provided in the file “Demographics of participants.xlsx”.

本数据集共收录1168例角膜炎患者的眼前节光学相干断层扫描（Anterior Segment Optical Coherence Tomography，简称AS-OCT）图像，其中包含768张全画幅图像（共来自6名患者）。每张图像均附带针对病灶与角膜的分割标注，全画幅图像还额外附带虹膜分割标注。该数据集按图像类别分为两个文件夹，分别命名为"Partial-frame_Dataset"与"Full-frame_Dataset"。在每个文件夹下，均设有"Original_AS-OCT_Images"子文件夹，其内存储格式为BMP的原始AS-OCT图像，文件命名规则为"n.bmp"；同时设有"Experts_Annotations"子文件夹，其内存储格式为JSON的对应标注文件，文件命名规则为"n.json"。参与者的匿名化人口统计学信息亦存储于"Demographics of participants.xlsx"文件中。

搜集汇总

背景与挑战

背景概述

该数据集包含1168张角膜炎患者的AS-OCT图像，分为部分帧和全帧两类，每张图像都有病变和角膜的分割标签，全帧图像还包括虹膜标注。数据以BMP格式存储，标注信息以JSON格式提供，适用于深度学习分割和3D重建研究。