zhaohandd/Unet-liverCT

Unet network for liver CT image segmentation

Unet-liverCT

Unet network for liver CT image segmentation
This work is based on U-Net: Convolutional Networks for Biomedical Image Segmentation AND unet. I develop the whole project to solve the problem of Unet network segmenting liver CT.

Overview

Data

The dataset 3D-IRCADb(3D Image Reconstruction for Comparison of Algorithm Database).This dataset includes 20 people's liver CT images, 15 of 20 have tumors.

Data Augmentation

I use keras.preprocessing.image to do the data augmentation in order to get enough images to train the network. You can do it or not.

Model Architecture

Train and Test

Dependencies

• python == 2.7.15
• tensorflow-gpu == 1.3.0
• keras == 2.0.5
  I did not test other versions, you can have a try.

Training

python trainUnet.py

Testing

python testUnet.py

Evaluating

python evaluate.py
The results of liver CT segmentation and tumor segmentation are based on the following indicators:Dice coefficient and RVD(relative volume difference) and VOE(volumetric overlap error).

Result