ImageSegmentation With Vnet3D

This is an example of the prostate in transversal T2-weighted MR images Segment from MICCAI Grand Challenge:Prostate MR Image Segmentation 2012

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、download trained data,download dataset：https://promise12.grand-challenge.org/download/ ,if you can't download it,i have shared it:https://pan.baidu.com/s/1y9YAAQKdD3OMOMyamx9MdA, password:whbf

2、the file of promise12Vnet3dImage.csv,is like this format:
D:\Data\PROMISE2012\Vnet3d_data\Vnet3d_patch_train\image/0_10
D:\Data\PROMISE2012\Vnet3d_data\Vnet3d_patch_train\image/0_11
D:\Data\PROMISE2012\Vnet3d_data\Vnet3d_patch_train\image/0_12
......
if you trained data path is not D:\Data\PROMISE2012,you should change the csv file path just like this:using C:\Data\ replace D:\Data\PROMISE2012.

3、when data is prepared,just run the vnet3d_train_predict.py

4、training the model on the GTX1080,it take 40 hours,and i also attach the trained model in the project,you also just use the vnet3d_train_predict.py file to predict,and get the segmentation result.

5、download trained model:https://pan.baidu.com/s/1kQ1SCVuBK6xJFR7cyKN7XQ password:0ytv

6、download test data: https://pan.baidu.com/s/1pDCQzTxUmyYdwDinBJKTuA, password:s0jt

Result

MICCAI Grand Challenge Result

the trained loss result

the Vnet3D model

the trained process:0 epoch——GTMask and PredictMask

1000 epoch——GTMask and PredictMask

10000 epoch——GTMask and PredictMask

the predict result