心房纤维化瘢痕挑战数据集

数据集获取（Figshare）

• 源站标题：CDEMRIS fibrosis scar challenge dataset
• DOI：10.6084/m9.figshare.4214532.v5
• 发布日期：2016-11-09
• 文件数：5 个
• 文件总体积：1.05 GB
• 源站页面：https://figshare.com/articles/dataset/CDEMRIS_fibrosis_scar_challenge_data_2012/4214532

命令行下载

curl -L -o "post.zip" "https://ndownloader.figshare.com/files/6875505"
curl -L -o "pre.zip" "https://ndownloader.figshare.com/files/6875508"
curl -L -o "sample_segmentation.zip" "https://ndownloader.figshare.com/files/6875511"
curl -L -o "LA_with_scar.png" "https://ndownloader.figshare.com/files/6875559"
curl -L -o "READ_ME.txt" "https://ndownloader.figshare.com/files/6878985"

文件清单

• post.zip（533.19 MB）
• pre.zip（487.05 MB）
• sample_segmentation.zip（26.34 MB）
• LA_with_scar.png（121.94 KB）
• READ_ME.txt（981 B）

文件总体积 1.05 GB（源站 API 实测）。

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。

数据说明

The Dataset

The dataset includes Late Gadolinium enhancement (LGE) cardiovascular magnetic resonance (CMR) imaging used to visualise regions of fibrosis and scarring in the left atrium (LA) myocardium. 
This can be important for treatment stratification of patients with atrial fibrillation (AF) and for assessment of treatment after radio frequency catheter ablation (RFCA). 
 The datasets included are from a segmentation challenge where the aim was to create a benchmarking framework for algorithms classifying fibrosis and scar from LGE CMR images. 
Dataset contents

The dataset includes two subsets called pre and post. These refer to the time-points when the images were acquired. The pre data contains pre-ablation imaging of fibrosis. The ‘‘post’’ data contains the post-ablation imaging data. There are 30 datasets in each pre and post folders. 

The folders are arranged as pXX, where XX is the case  number. Within each folder you will find two images in NRRD format. 

These images follow a naming convention: 

• de_xx_nn

• la_seg_xx_nn

• The ‘‘xx’’ can be ‘‘a’’ or ‘‘b’’. The ‘‘a’’ refers to the pre-scan and the ‘‘b’’ refers to the post-scan.

•  The ‘‘nn’’ refers to the case number

Note the de_xx_nn refers to the delayed-enhancement (DE) scan. And the la_seg_xx_nn refers to the binary segmentation of the left atrial endocardium. The segmentation is registered to the delayed-enhancement scan. The images will generally be of varying quality. 
 

References

Karim et. al. Evaluation of current algorithms for segmentation of scar tissue from late Gadolinium enhancement cardiovascular magnetic resonance of the left atrium in Journal of Cardiovascular Magnetic Resonance, 15(105), 2013. Link

许可

CC BY 4.0

引用

学术使用请引用 DOI 10.6084/m9.figshare.4214532.v5（source: https://figshare.com/articles/dataset/CDEMRIS_fibrosis_scar_challenge_data_2012/4214532）。