This GHMM_nD package is developed by Chao Huang and Hongtu Zhu from the BIG-S2 lab and Marc Niethammer from the uncbiag lab.
Gaussian hidden Markov model (nD) is a Matlab based package for statistical nD (n=2,3) imaging spatial heterogeneity analysis. A Gaussian hidden Markov model is introduced to build the spatial heterogeneity of imaging intensity across different patients. The statistical inference results are used in diseased region detection in both personal level and populational level.
We provide a command script (GHMM_nD_scp.m) to run the data analysis with GHMM_nD. Please see the file GHMM_nD_scp.m for usage. We also provide some simulated 3D imaging dataset for testing the package.
This toolbox is written in a user-friendly GUI interface. Here are some explainations of the interface:
- Load raw data
In this panel, you need to select (a). the folder where mask data (mask file, mat format) is stored; (b). the file containing the covariate matrix (could be either txt format or mat format); (c). the folder where imaging dataset (mat format) is stored
- Clear & Run
In this panel, you need to select the folder where to save all the output files. The 'Clear' button help you to clear all the preselcted input information, and the 'Run' button is clicked when all the input information is selected and the analysis is ready to go.
- Export Results
In this panel, you can plot the results you are interested in.