CSDI白内障诊断数据集
该数据集包含187张经过专家标注的眼底图像，每张图像均提供白内障严重程度评分和双语诊断描述，属于医学影像与文本多模态数据，主要用于支持自动化白内障筛查、疾病分级以及眼底图像解释等医学人工智能研究方向。
基本信息
资源简介
CSDI白内障诊断数据集
【数据集背景】
本数据集源站标题为《rainyNighti/CSDI》。源站首次发布：2025-07-22。
【数据内容】
源站原始描述：Fundus Image Dataset for Fine-Grained Cataract Severity Staging and Diagnostic Captioning # CSDI: A Fine-Grained Fundus Image Dataset of Cataract Severity and Diagnostic Images The **CSDI Cataract Diagnosis Dataset** is a curated collection of 187 fundus images with expert annotations, including cataract severity scores and bilingual diagnostic descriptions (English and Chinese). This dataset supports research in automated cataract screening, grading, and fundus image interpretation using both i…
【获取与许可】
源站页面：https://github.com/rainyNighti/CSDI
使用许可：NOASSERTION。
【使用注意】
· 源站最后修改于 2026-08-07，引用时请注意版本时效。
下载信息
注册下载
需要注册 Kaggle 账号并登录后下载，适合需要跟踪下载记录和使用 API 的用户。
暂未开放公开下载
无需注册即可直接获取公开样本或文档，适合快速预览和评估数据集质量。
免登录有偿下载
公开数据集受托下载与技术交付服务。
提供高速下载与技术交付服务（收技术服务费，非数据销售）
暂未开放当前数据集主要来源为 Kaggle 公开托管，完整影像包建议通过原始链接或 Kaggle API 获取。
使用方式
数据集获取
git clone https://github.com/rainyNighti/CSDI.git
curl -L -o repo.zip https://github.com/rainyNighti/CSDI/archive/refs/heads/main.zip
unzip repo.zip
源站 README 摘录（使用方式）
CSDI: A Fine-Grained Fundus Image Dataset of Cataract Severity and Diagnostic Images
The CSDI Cataract Diagnosis Dataset is a curated collection of 187 fundus images with expert annotations, including cataract severity scores and bilingual diagnostic descriptions (English and Chinese). This dataset supports research in automated cataract screening, grading, and fundus image interpretation using both image and text modalities.
This dataset is licensed under CC BY-NC 4.0. See LICENSE.md for details.
📁 Dataset Overview
- Total images: 187
- Image formats:
.pngand.jpg - Annotation file:
CSDI_annotations.csv - Labels: Cataract severity score (0–10), expert-written diagnoses, optic disc localization, optic disc clarity
📂 File Structure
CSDI/
├── csdi/ # Training and evaluation code
├── original_images/ # Folder containing 187 fundus images (.png and .jpg)
│ ├── cataract_001.png
│ ├── cataract_002.jpg
│ └── ...
├── CSDI_annotations.csv # Annotation CSV file (UTF-8 encoded)
├── README.md # Dataset description
├── LICENSE.md # Dataset license (CC BY 4.0)
├── crop_fundus_images.py # Script for cropping fundus images
├── augment_fundus_images.py # Script for image augmentation
├── dataset_split.py # Includes information about the validation data split
├── prompt.py # Prompts for training and inference
└── create_dataset_json.py # Create training and inference JSON files
🗒️ CSV Annotation File (CSDI_annotations.csv)
- The CSV annotation file is encoded in UTF-8 to ensure proper display of English and Chinese characters.
| Column Name | Description |
|---|---|
id |
Image file name (e.g., cataract_001.png) |
score |
Cataract severity score (range 0–10) |
English_diagnosis |
Diagnosis in English |
Chinese_diagnosis |
Diagnosis in Chinese |
original_image_width |
Original image width in pixels |
original_image_height |
Original image height in pixels |
fundus_region_x1 / y1 / x2 / y2 |
Coordinates of the annotated fundus region in pixels; -1 if not visible |
optic_disc_clear |
Optic disc visibility (visible / not visible) |
| `optic_disc_x1 |
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。
数据缺失？
依托陕西、四川两大基地，我们与超过十家三甲医院建立直接合作关系，覆盖合规授权、采集治理、专业标注、数据交付的全流程，为AI医疗团队提供即拿即用的高质量临床数据。
⚡️ 需要数据支持或标注服务？立即联系我们获取专业方案。
