BenchReAD

数据集获取

git clone https://github.com/DopamineLcy/BenchReAD.git

curl -L -o repo.zip https://github.com/DopamineLcy/BenchReAD/archive/refs/heads/main.zip
unzip repo.zip

源站 README 摘录（使用方式）

Getting started

1 Requirements

OS: Ubuntu 20.04 LTS.

Language: Python 3.10.8

Create a virtual environment and install the dependencies through the following command:

pip install numpy
pip install pandas
pip install tqdm
pip install scikit-image
pip install opencv-python

2 Building fundus benchmark

2.1 Training and validation sets establishment

The traning and validation sets of the fundus benchmark are build on the EDDFS [1] and BRSET [2] datasets. The following steps are used to establish the training and validation sets.

1. Download the EDDFS and BRSET datasets from the links.

2. Put the EDDFS data in the ./data/EDDFS/OriginalImages folder and the annotations in the ./data/EDDFS folder.

3. Preprocessing the EDDFS data using the EDDFS_preprocessing.ipynb and EDDFS_split.ipynb notebooks.

4. Put the BRSET data in the ./data/brazilian-ophthalmological/ folder.

5. Preprocessing the BRSET data using the BRSET_preprocessing.ipynb and BRSET_split.ipynb notebooks.

2.2 Testing sets establishment

The testing sets of the fundus benchmark are build on the RIADD [3] and JSIEC [4] datasets. The following steps are used to establish the testing sets.

1. Download the RIADD and JSIEC datasets from the links.

2. Put the RIADD data in the ./data/RIADD/ folder.

3. Preprocessing the RIADD data using the RIADD_preprocessing.ipynb and RIADD_split.ipynb notebooks.

4. Put the JSIEC data in the ./data/JSIEC/ folder.

5. Preprocessing the JSIEC data using the JSIEC.ipynb notebook.

3 Building OCT benchmark

3.1 Training & validation set establishment

The training and validation sets of the OCT benchmark are build on the OCT 2017 [5] dataset. The following steps are used to establish the training and validation sets.

1. Download the OCT 2017 data

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/DopamineLcy/BenchReAD