Go Back to the Main Dissemination Page..


These programs are developed in house for processing MRI images and other MRI related concerns. All programs are provided as-is and should be considered works in progress. Please email the respective authors with comments, complaints, and errors.

Contents

edit

SequenceTree

SequenceTree is an open-source pulse programming environment with a large user community which is growing every day.

Many pulse sequences are available for download, including:

Spoiled Gradient Echo, Balanced Gradient Echo, Fast Spin Echo, Diffusion, T1ρ.

To learn more about SequenceTree or join the SequenceTree community, visit thesouthpoles.com/sequencetree



edit

Source Code

Source code for projects is available on Bloch or Subversion.

edit

2D Mutual Information Image Co-registration (MATLAB)

2D Mutual Information Image Co-registration (MATLAB)

Author: Matthew Sochor

Co-registers two images through rotation and translation in 2 dimensions. See the above link for more documentation. Code is written in c++ and compiled in MATLAB.

edit

3D Volume Exponential Fit (MATLAB)

Bloch: \Image Processing Programs\t1rrsq-2007b

Authors: Walter Witschey, Matthew Sochor

This function fits an input 4D double matrix (N 3D Volumes) to an exponential decay model using linear least squares regression on a pixel by pixel basis. The code is written in c++ and compiled using the MATLAB compiler. Optional output includes R squared maps and/or Chi squared maps. 

Exponential decay model: S(t) = S0 * exp( − t / Tm)

Syntax: [Exp_map (R_squared_map)] = t1rmap(Times, Input_matrix);

Input
Times: N x 1 matrix containing the time points along the exponential curve when the measurements are taken. Parameter t in the above formula.
Input_matrix: N x X x Y x Z double matrix. N is the number of samples. X,Y, and Z are the input matrix dimensions. Parameter S(t) in the above formula.
Output
Exp_map: X x Y x Z double matrix. Parameter T_m in the above formula.
R_squared_map: (optional) X x Y x Z double matrix. R squared fit parameter for each pixel. Measures the accuracy of the fit.
edit

Sorting DICOM images (MATLAB and IDL)

Bloch: \Image Processing Programs\01 - standalone_sort - IDL\standalone_sort.pro

Author: Mark Elliott

Use this program to sort knee study images

Bloch: \Image Processing Programs\standalone_sort.m

Author: Matthew Sochor

Use this program to sort images from conquest Dicom server

Bloch: \Image Processing Programs\sortconquest.m

Author: Matthew Sochor

Use this program to sort CERND study images specifically. Calls standalone_sort.m

edit

coreg_viewnix_v1 (MATLAB GUI/3DViewnix)

Bloch \Image Processing Programs\coreg_viewnix_v1\

Author: Matthew Sochor, Matthew Fenty

This program performs all the initial steps to take folders of sorted DICOMS from the knee studies and have them become useful data.

It inputs everything into matlab, allows for manual alignment of anatomicals to the t1rho images, co-registers everything using 3DViewnix (NOTE: this must be installed correctly!!) and then makes T1rho and/or T2 maps, saves everything as .MAT space and sets up folders for segmentation.

Segment in 3DViewnix and open the output masks and .MAT space in mmrrcc_study_stats

edit

MMRRCC Study Stats (MATLAB GUI)

Subversion: http://www.thesouthpoles.com/mmrrcc_study_stats

Author: Matthew Sochor, Matthew Fenty

This program is used to process knee study data for current knee studies. It takes in .MAT spaces from the Study_viewnix_v# series of programs along with .png masks made with 3DViewnix. The program subdivides axial and coronal images into medial and lateral and then into superficial, middle and deep layers (see internal_eroder). It then generates statistical data for each ROI.

edit

MMRRCC Study Viewer

Subversion: http://www.thesouthpoles.com/mmrrcc_study_viewer

Author: Matthew Sochor, Walter Witschey

This is a Matlab program for viewing knee studies, measuring T1rho/T2 values of segmented images, and quickly generating publication worthy color overlay images.

This program takes in .MAT spaces generated from MMRRCC Study Stats.

edit

Internal Eroder (MATLAB)

Bloch: \Image Processing Programs\internal_eroder.m

Author: Matthew Sochor

This program subdivides an ROI into three vertically separated subregions. The goal is to potentially measure any "layered" characteristics of cartilage as has been noted in T1rho and T2 quantitative MRI.

edit

CERND_v# (MATLAB GUI)

Bloch: \Image Processing Programs\CERND - 2D TSE\cernd_v5.m

Author: Matthew Sochor

This program takes in CERND protocol MRI 2D TSE T1rho images and MP-RAGE anatomical images. It co-registers the T1rho images to the earliest TSL image using 2D MI coregistration and generates a T1rho and R squared map. ROIs are generated using HAMMER and are aligned using the MPR anatomicals. ROI statistical data is saved.

edit

AOSpine Lumbar Spine Analysis (MATLAB GUI)

Bloch: \Image Processing Programs\Degen_do_not_edit_that_means_you_sochor\Degen_gui.m

Author: Matthew Fenty

This program takes in DEGEN protocol 2D FSE 2D Proton Density images and raw T1r image k-space. It takes the raw multi-channel k-space data and reconstructs individual spin-lock images as well as generates a T1r map. Segmentation is accomplished which then allows for statistical analysis in the same package.

edit

distribute_raw_data (MATLAB)

Bloch: \Image Processing Programs\distribute_raw_data (MATLAB)

Author: Matthew Sochor, Jeremy Magland

This program takes in meas.out data, applies a specific ST template.tmp file, and writes ADC0/1.mda files. Performs identical to the same named function in chainlink, however this can be compile for use directly in MATLAB. Read .mda files using readmda.m (located in the same folder on Bloch).

edit

navigators

Bloch: \Image Processing Programs\navigators

Author: Matthew Sochor

This program minimizes the least squares difference between two profiles. Used for translation navigator correction. See renavigate.m for use in correcting kspace data with this program.

edit

t1r_3dbrain recon

Bloch: \Image Processing Programs\t1rbrain_recon

Author: Matthew Sochor, Walter Witschey, Jeremy Magland

Giant conglomeration program that is adaptable and useable for reconning any sequence tree meas.out space with all kinds of funky kspace adquisitions, navigators, and postprocessing steps. Output is a t1rmap, but you could make it DTI or a sodium map if you like.

Retrieved from "http://www.mmrrcc.upenn.edu/mediawiki/index.php/Programs"
Last modified June 4, 2008 12:49 pm / Skin by Kevin Hughes
MediaWiki