junqiangchen/LiTS---Liver-Tumor-Segmentation-Challenge 开源项目中文解读：技术栈、适用场景与上手建议

258

MIT License

项目介绍

junqiangchen/LiTS---Liver-Tumor-Segmentation-Challenge 是一个托管在 GitHub 上的开源项目，主要使用 Python 编写，采用 MIT License 许可证，目前获得 258 个 Star、76 次 Fork。

仓库全名: junqiangchen/LiTS---Liver-Tumor-Segmentation-Challenge
开源许可证: MIT License
Star 数: 258
Fork 数: 76

README 摘要

ImageSegmentation With Vnet3D This is an example of the CT images Segment from LiTS---Liver-Tumor-Segmentation-Challenge Prerequisities The following dependencies are needed: numpy >= 1.11.1 SimpleITK >=1.0.1 opencv-python >=3.3.0 tensorflow-gpu ==1.8.0 pandas >=0.20.1 scikit-learn >= 0.17.1 How to Use (re)implemented the model with tensorflow in the paper of "Milletari, F., Navab, N., & Ahmadi, S. A. (2016) V-net…

完整内容见 GitHub 原始 README 或本页。

常见问题

LiTS---Liver-Tumor-Segmentation-Challenge 用什么语言开发？

主要使用 Python 编写，采用 MIT License 许可证，目前有 258 个 Star。

LiTS---Liver-Tumor-Segmentation-Challenge 可以免费商用吗？

如何获取 LiTS---Liver-Tumor-Segmentation-Challenge 的源码？

可以直接克隆仓库：git clone https://github.com/junqiangchen/LiTS---Liver-Tumor-Segmentation-Challenge.git，默认分支为 master；安装与运行步骤见仓库 README。

LiTS---Liver-Tumor-Segmentation-Challenge 最近还在维护吗？

仓库最近一次更新为 2021年8月27日，当前没有未关闭的 Issue；Star 数为 258，可作为社区活跃度的参考。

关于

LiTS - Liver Tumor Segmentation Challenge

描述取自 GitHub 官方 API，中文内容由 GitRead 中文社区编辑团队整理。

简要分析

热度分

422

价值分

178

活跃状态

活跃

tensorflow liver segmentation vnet liver-segment numpy cv2 os simpleitk

GitHub

语言占比

Python

贡献者

junqiangchen

Dxye

README

ImageSegmentation With Vnet3D

This is an example of the CT images Segment from LiTS—Liver-Tumor-Segmentation-Challenge

Prerequisities

The following dependencies are needed:

numpy >= 1.11.1
SimpleITK >=1.0.1
opencv-python >=3.3.0
tensorflow-gpu ==1.8.0
pandas >=0.20.1
scikit-learn >= 0.17.1

How to Use

(re)implemented the model with tensorflow in the paper of “Milletari, F., Navab, N., & Ahmadi, S. A. (2016) V-net: Fully convolutional neural networks for volumetric medical image segmentation.3DV 2016”

1、Preprocess

LiTS data of image and mask are all type of .nii files,in order to train and visulise,convert .nii file to .bmp file.
Liver data preparing,i have tried many patch size,and finally using the patch(256,256,16),if you have better GPU,you can change 16 to 24 or 32:run the getPatchImageAndMask.py
Tumor data preparing,using the patch(256,256,16):run the getPatchImageAndMask.py,disable the line gen_image_mask(srcimg, seg_liverimage, i, shape=(16, 256, 256), numberxy=5, numberz=10) and enable the line gen_image_mask(srcimg, seg_tumorimage, i, shape=(16, 256, 256), numberxy=5, numberz=10),and change the trainLiverMask to trainTumorMask
last save all the data folder path into csv file: run the utils.py

the file like this:

G:\Data\segmentation\Image/0_161

G:\Data\segmentation\Image/0_162

G:\Data\segmentation\Image/0_163

2、Liver and Tumor Segmentation

the VNet model

train and predict in the script of vnet3d_train.py and vnet3d_predict.py

3、download resource

liver segmentation trained model,log,test data can download on here:https://pan.baidu.com/s/1ijK6BG3vZM4nHwZ6S2yFiw, password：74j5
LiTS data have 130 cases,using 0-110 cases trainging,and other is testing.testing result can download on here:https://pan.baidu.com/s/1A_-u7tJcn7rIqnrLaSqi4A password：22es
LiTS train and test source data can download here:https://pan.baidu.com/s/1-kxJ7reS4kq5ypitfmQxeg password：nlrd

Result

Trained Loss

Liver Segment Result

Liver leaderboard

test case segmentation result can see in the file of 35.mp4,38.mp4 and 51.mp4

first col is srcimage,second col is GroundTruth Mask image,third col is VNet segmentation image

Lesion leaderboard

Contact

https://github.com/junqiangchen
email: [email protected],[email protected],[email protected]
Contact:junqiangChen,dexianYe,xingTao
WeChat Public number: 最新医学影像技术

LiTS---Liver-Tumor-Segmentation-Challenge