Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.
You can not select more than 25 topics
Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
huangweijian
3175f3b68d
|
2 years ago | |
|---|---|---|
| .gitignore | 2 years ago | |
| D-Unet.png | 2 years ago | |
| LICENSE | 2 years ago | |
| README.md | 2 years ago | |
| Statistics.py | 2 years ago | |
| data_load.py | 2 years ago | |
| model.py | 2 years ago | |
README.md
D-UNet
D-UNet: a dimension-fusion U shape network for chronic stroke lesion segmentation
Author
Yongjin Zhou, Weijian Huang, Pei Dong, Yong Xia, and Shanshan Wang.
项目简介
1. 功能
采用D-UNet实现对ATLAS数据集的图像分割，兼顾了3D特征提取及高效的实现。
2. 性能
| DSC | Recall | Precision | Total parameters |
|---|---|---|---|
| 0.5349±0.2763 | 0.5243±0.2910 | 0.6331±0.2958 | 8,640,163 |
3. 评估指标
DSC(Dice Similarity Coefficient)、Recall、Precision
4. 使用数据集
ATLAS（Anatomical Tracings of Lesions-After-Stroke dataset）
[1] Liew, Sook-Lei, et al. "A large, open source dataset of stroke anatomical brain images and manual lesion segmentations." Scientific data 5 (2018): 180011.
运行环境与依赖
代码运行的环境与依赖。如下所示：
| 名称 | 版本 |
|---|---|
| ubuntu | 16.04 |
| Tensorflow | 1.10.0 |
| Keras | 2.2.0 |
| Python | 3.6.0 |
输入与输出
代码的输入与输出。如下所示：
| 名称 | 说明 |
|---|---|
| 输入 | 3D灰度图像，大小为192X192X4X1（宽x高x切片数x通道）的连续切片。 |
| 输出 | 分割结果。0表示背景，1表示脑卒中区域 |
运行方式
python Stroke_segment.py
