CINCH

From VISTA LAB WIKI

Jump to: navigation, search

CINCH is a tool for visualization and segmentation of estimated white matter fibers (tracts). The program was developed by David Akers' in collaboration with our team. Dave designed an innovative interface that is gesture based, not ROI based. This makes fiber segmentation easier and more fun than ROI drawing.

CINCH is a C++ application. It is designed to work either independently or in coordination with mrDiffusion. You will need around 2GB of memory to run it. CINCH runs under Windows and Linux.

Contents

[edit] Installing

CINCH Interface

For all platforms, you can begin by downloading a sample data.

Using that data set and the executable below, you can follow the CINCH Tutorial.

[edit] Windows

To install CINCH on a Windows machine

  1. Download the CINCH_win32.zip file from our download page.
  2. Unzip the files to a directory of your choosing. The executable will be named CINCH.exe. You will either find it in the zip file or in the /CINCH/CINCH.exe.
  3. Run the CINCH.exe executable.
  • If you plan to use CINCH together with mrDiffusion, the executable (CINCH.exe) needs to be in your Matlab path. Add the path to the executable to your startup file. CINCH can live pretty much anywhere, for example, C:\CINCH\. Locating it under Program Files or any directory containing a space in its name might be too much of a challenge for Matlab. Sigh.

Test if CINCH was added to your Matlab path by typing "which 'CINCH.exe'" or "which 'CINCH.glx'" into the Matlab prompt.

[edit] Linux

CINCH has been compiled for linux 32-bit and 64-bit platforms. Those versions are available on the software page.

  1. Download the 32- or 64-bit version of CINCH. If you're not sure if your system is 32-bit or 64-bit try running the following from the command line: file /sbin/init
  2. Extract the CINCH folder contents to your system (e.g., /usr/local/CINCH64/CINCH).
  3. You can now run the CINCH executable via the command line by changing to the directory into which you've extracted CINCH and typing: ./CINCH

If you try to run CINCH you get an error message, you may need to install a couple of libraries.

  • libvtk
  • libtiff

To install these libraries you should run your installer and search for libvtk and libtiff. You will need sudo access.

If you're running ubuntu, from the command line:

> sudo synaptic

In synaptic search for libvtk. Install the version of libvtk that is followed by a version number. You shouldn't need the dev or qt versions. It should be something like libvtk5 for a 32bit ubuntu.


If you're running Fedora, the following should work from the command line:

> sudo yum install libvtk
  • Repeat the same process for libtiff

[edit] Keyboard shortcuts

Below are the keyboard/mouse shortcuts, for convenience:

Mouse commands

  • Rotate the scene - Click and drag left mouse
  • Draw a mark - Hold down Ctrl, Click and drag left mouse
  • Zoom/spin - Click and drag right mouse
  • Pan - Click and drag middle mouse

General

  • UndoCtrl-Z
  • RedoCtrl-Y

Tomograms

  • Select a tomogramA,S,C (axial, sagittal, coronal)
  • Toggle tomogram visibilityShift-A, Shift-S, Shift-C (axial, sagittal, coronal)
  • Slide current tomogram<, >, mouse wheel
  • Reposition tomogramsDouble click on any visible tomogram
  • Cycle tomogram imageCtrl-B

Pathways

  • Select a fiber group1-8 number keys, or click on a group
  • Toggle fiber group visibilityShift-[#], or click on already-selected group
  • Thicken fiber tractsw
  • Select all pathwaysCtrl-A
  • Deselect all pathwaysShift-Ctrl-A
  • Grow/shrink a selectionplus/minus keys
  • Lock/unlock pathway assignmentsCtrl-K

File I/O

  • Load interface stateCtrl-L
  • Save interface stateCtrl-S: cst file (with distinct fiber groups)
  • Load pathwaysCtrl-H: pdb file
  • Load distance matrixCtrl-M

[edit] Developers

CINCH was originally written by David Akers. It has been updated for mrVista integration by Shireesh Aggrawal, Tony Sherbondy and Bob Dougherty. Source is available for developers through the SVN repository.


[edit] Windows 64-bit issues

Dave Akers is working on resolving 64-bit windows-related issues on the next release of CINCH, but there are some hacks for the meantime. These hacks may not be necessary if you can run a version of MATLAB in 32-bit mode (e.g., MATALBR2006b), or if you are working with older versions of mrDiffusion and DTI data. However, if you are working with the current version of mrDiffusion, and DTI data in nifti format, you will need to do the following in order for CINCH to work.

  1. In the subject's bin directory, create a directory called "backgrounds"
  2. In this directory, copy over the backgrounds you'd like to use (e.g,. 't1.nii.gz'). You can try to make a softlink rather than a duplicate, but it didn't work on my machine.
  3. In MATLAB, run the following example commands (changing the nifti file name as appropriate):
  4.     ni=readFileNifti('t1.nii.gz');
        ni.data=mrAnatHistogramClip(double(ni.data),0.4,0.99);
        ni.data=single(ni.data);
        ni.fname='bin/backgrounds/t1.nii.gz'; % this is where it will save the file, alter the path as necessary
        writeFileNifti(ni);
    
  5. Start CINCH by doubleclicking on CINCH.exe
  6. From CINCH: File --> Set Subject Datapath: navigate to the subject's directory and select the bin directory
  7. From dtiFiberUI: View --> CINCH --> Push current paths

[edit] Linux 64-bit tips

To access CINCH from a 64-bit linux machine, type the following command:

  /usr/local/CINCH64/CINCH

If CINCH is not called from the correct directory, it will not be able to find the .png images used for the "Touch", "Match", etc buttons.

Personal tools