DoB24/fundus-10class-augmented

Fundus 10-Class Augmented Dataset是一个医学影像数据集，专门用于眼底图像分类。它包含10种疾病类别：中心性浆液性脉络膜视网膜病变、糖尿病视网膜病变、视盘水肿、青光眼、健康、黄斑瘢痕、近视、翼状胬肉、视网膜脱离和视网膜色素变性。数据集共有5,335张原始图像和16,242张增强图像（通过旋转、翻转、亮度/对比度抖动和高斯模糊生成）。采用组感知测试分割策略，确保增强图像与原始图像分组，避免患者内数据泄漏。图像尺寸可变，在训练时通常调整为224×224或299×299。数据来源于Mendeley数据集，遵循CC BY 4.0许可证。该数据集用于深度学习模型训练，特别是眼底病变分类任务。The Fundus 10-Class Augmented Dataset is a medical imaging dataset for fundus image classification. It includes 10 disease classes: Central Serous Chorioretinopathy, Diabetic Retinopathy, Disc Edema, Glaucoma, Healthy, Macular Scar, Myopia, Pterygium, Retinal Detachment, and Retinitis Pigmentosa. The dataset consists of 5,335 original images and 16,242 augmented images generated through rotation, flip, brightness/contrast jitter, and Gaussian blur. A group-aware test split is employed to confine augmented children of each original image to a single split, eliminating intra-patient leakage. Image sizes are variable and resized to 224×224 or 299×299 during training. The data is sourced from the Mendeley dataset under the CC BY 4.0 license. This dataset is used for training deep learning models, particularly for retinal disease classification tasks.