Skip to content

Codes that I have written to complete promise12 prostate segmentation competition.

Notifications You must be signed in to change notification settings

mirzaevinom/promise12_segmentation

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 

Repository files navigation

PROMISE12 Challenge - Automated Segmentation of Prostate Structures from MR Images

This is Keras implementation of a fully convolutional neural network with residual connections for automatic segmentation of prostate structures from MR images.

More info on this competition can be found on Grand Challenges website. Data can be downloaded from https://promise12.grand-challenge.org/download/

The network architecture was inspired by U-Net: Convolutional Networks for Biomedical Image Segmentation and by Keras implementation of the model by Paul-Louis Pröve.

The predictions of this model achieved the score 83.70 and was ranked #8 in the competition. For more details about the model and the implementation see the file project_summary.pdf

images/schematic.png

How to use

Dependencies

This tutorial depends on the following libraries:

  • scikit-image, numpy, matplotlib, scipy
  • SimpleITK
  • OpenCV
  • Tensorflow >=1.4
  • Keras >= 2.0

This code should also be compatible with Theano backend of Keras, but in my experience Theano is slower than TensorFlow.

Running the model

  • Run python train.py to pre-process the data and train the model. Model weights are save in file ../data/weights.h5 .

  • Run python test.py to test the model on the train and validation set and generate some images with some best and worst predictions.

About

Codes that I have written to complete promise12 prostate segmentation competition.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages