数据集获取

git clone https://github.com/TheBeastCoding/glaucoma-dataset-metadata.git

curl -L -o repo.zip https://github.com/TheBeastCoding/glaucoma-dataset-metadata/archive/refs/heads/main.zip
unzip repo.zip

源站 README 摘录（使用方式）

New Updates

• 01/01/2026 : Adding new dataset, Hillel Yaffe Glaucoma Dataset (HYGD), https://physionet.org/content/hillel-yaffe-glaucoma-dataset/1.0.0/ and Fundus-AVSeg [https://figshare.com/articles/dataset/Fundus-AVSeg/27938034?file=54093641]
• 08/03/24 : A new abstract is available to the public, introducing a novel approach to glaucoma detection: Assessment of Retinal Vasculature for Glaucoma Detection: A Comparative Analysis of Human Expertise and Deep Learning Algorithms.
• 05/15/24 : The EyePACS-light-V2 code folder on Kaggle has a new PyTorch template for quick and easy glaucoma detection setup. This model uses the lightweight MobileNetV3 and achieved a test accuracy of 92.6%: https://www.kaggle.com/code/deathtrooper/pytorch-easy-setup-for-glaucoma-detection-92-6
• 03/09/24 : EyePACS-light-V2 now has a 10.0 Kaggle usability score: supplemental metadata.csv file added to dataset.
• 01/20/24 : EyePACS-light-V2 preprocess high-level overview is now available on the Kaggle dataset link in the about section (scroll all the way down) if you are curious on how the dataset was derived
• 12/28/23 : EyePACS-light-V2 94.94% test accuracy benchmark using ConvNeXtTiny: https://www.kaggle.com/code/deathtrooper/benchmark-94-94-convnexttiny
• 12/12/23 : EyePACS-light-V2 is HERE!!! Be the first to benchmark your model with this improved dataset! Download from kaggle: https://www.kaggle.com/datasets/deathtrooper/glaucoma-dataset-eyepacs-airogs-light-v2/data

Citation Note

If you found this catalog helpful, please consider citing the following:

• Riley Kiefer, Muhammad Abid, Jessica Steen, Mahsa Raeisi Ardali, and Ehsan Amjadian. 2023. A Catalog of Public Glaucoma Datasets for Machine Learning Applications: A detailed description and analysis of public glaucoma datasets available to machine learning engineers tackling glaucoma-related problems using retinal fundus images and OCT images. In Proceedings of the 2023 7th International Conference on Information System and Data Mining (ICISDM ''23). Association for Computing Machinery, New York, NY, USA, 24–31. https://doi.org/10.1145/3603765.3603779

Public Glaucoma Dataset Catalog

[Help expand this repository by providing links/publications to new glaucoma datasets!]

Repository Table of Contents

• README.md : Glaucoma overview, relevant research, and dataset access links
• benchmark-eyepacs-airogs-light.md : Leader

数据加载示例（图像类）

from PIL import Image
import glob, os

files = (glob.glob(os.path.join(path, "**", "*.png"), recursive=True)
       + glob.glob(os.path.join(path, "**", "*.jpg"), recursive=True)
       + glob.glob(os.path.join(path, "**", "*.tif"), recursive=True))
print("图像文件数:", len(files))
img = Image.open(files[0]); print("尺寸/模式:", img.size, img.mode)

# torchvision Dataset 方式:
# from torchvision import datasets
# ds = datasets.ImageFolder(path)  # 要求 子目录=类别

目录组织与标注格式以源站说明和下载后实际文件为准。

完整仓库：github.com/TheBeastCoding/glaucoma-dataset-metadata