CC1.net真菌性角膜炎病例数据集

该数据集源自一项针对真菌性角膜炎（Fungal Keratitis, FK）的多中心研究，包含2015年4月至2022年6月期间从中国三个城市收集的临床数据。所有病例均通过角膜培养或角膜活检确诊。研究排除了年龄超过65岁、患有糖尿病、自身免疫性疾病、甲状腺相关疾病或妊娠的患者，同时排除了经培养证实为细菌、阿米巴或其他感染的患者。最终共纳入355例患者，其中广东省中山眼科中心212例、陕西省西安市第一医院83例、广西壮族自治区人民医院60例。
由5名持证眼科医生采用标准流程对图像质量进行评估，共筛选出8,697张高质量图像。这些图像进一步划分为三个子数据集：
分类数据集（共7,086张图像）：训练集4,255张、验证集1,413张、测试集1,418张；
分割数据集（共56张图像）：训练集46张、测试集10张；
外部验证集（共1,555张图像）：其中1,455张用于分类模型验证，100张用于分割模型验证。

This dataset is derived from a multicenter study on fungal keratitis (FK), containing clinical data collected from three cities in China between April 2015 and June 2022. All cases were confirmed via corneal culture or corneal biopsy. The study excluded patients aged over 65 years, those with diabetes, autoimmune diseases, thyroid-related disorders, or pregnancy, as well as patients with culture-proven bacterial, amoebic, or other infections. A total of 355 patients were finally included, with 212 cases from Zhongshan Ophthalmic Center in Guangdong Province, 83 cases from Xi'an First Hospital in Shaanxi Province, and 60 cases from People's Hospital of Guangxi Zhuang Autonomous Region.

Five certified ophthalmologists evaluated the image quality using a standard protocol, and a total of 8,697 high-quality images were screened out. These images were further divided into three subsets:
- Classification dataset (7,086 images in total): 4,255 for training, 1,413 for validation, and 1,418 for testing;
- Segmentation dataset (56 images in total): 46 for training, 10 for testing;
- External validation dataset (1,555 images in total): 1,455 for classification model validation, and 100 for segmentation model validation.

搜集汇总

背景与挑战

背景概述

该数据集是一个真菌性角膜炎病例数据集，包含2015年4月至2022年6月期间从中国三个城市医院收集的355名患者数据，共8697张高质量图像，经过严格筛选和眼科医生分级。数据集划分为分类、分割和外部验证子集，专为人工智能在真菌性角膜炎诊断中的研究和模型评估设计，具有高可靠性和科学性。