This repository contains code and examples for the Spatial Heterogeneity Analysis by Recursive Partitioning (SHARP) algorithm. The algorithm recursively segments a tumor on 2D or 3D images and outputs the amount of heterogeneity at various distance scales.
Created by Michael Gensheimer and Andrew Trister at the University of Washington. Direct questions to michael.gensheimer@gmail.com
The code was tested in MATLAB R2012b in Linux. It requires the Image Processing Toolbox and Sameer Agarwal's free Spectral Clustering Toolbox, available here: http://homes.cs.washington.edu/~sagarwal/code.html
To calculate Haralick texture features (one of the comparison methods mentioned in the paper), install Avinash Uppuluri's GLCM_Features4.m, available here: http://www.mathworks.com/matlabcentral/fileexchange/22354-glcmfeatures4-m-vectorized-version-of-glcmfeatures1-m-with-code-changes
To install, simply add the code and toolbox directories to the MATLAB path.
Run mri.m for a demonstration using one slice from MRI of a breast tumor. Run synthetic.m for a demonstration using synthetic images.