About COMKAT R3.2
COMKAT R3.2 is released in April 2009. It is the latest release of COMKAT.
COMKAT on MATLAB
COMKAT Standalone Application (Beta)
For Windows, click here to download COMKAT R3.2 (Standalone version).
This is a beta version of COMKAT standalone application. Please provide feedback if you do find bugs and errors.
Other platforms will be provided recently.
NOTE: You must log in first to download.
OS: Mac OS, Linux and Windows.
MATLAB: As the mex files are compiled under different versions of MATLAB, COMKAT on different platforms requires different MATLAB release.
- Windows: MATLAB R2006a and above
- Windows 64-bit: MATLAB R2008a and above
- PowerPC-based Mac OS: Not supported any more.
- Intel-based Mac OS: MATLAB R2008a and above
- Linux: MATLAB R2008a and above
- Linux 64-bit: MATLAB R2008a and above
1. Parametric imaging capabilities
2. Distributed computing supports
3. Misc bug fix
4. Improved GUI appearance
5. Standalone distributions
For COMKAT Standalone application, just unzip the zip file and run 'COMKAT.exe' by double clicking it. You may ignore the following steps.
COMKAT on MATLAB
To install COMKAT to your system, follow these steps.
1. Download COMKAT_R3.2.zip from http://comkat.case.edu
2. Unzip it to your computer.
3. Set up search path for COMKAT. You may do either of the following:
(1) Automatic: Open MATLAB, run 'SetComkatPath.m' under the COMKAT root folder.
(2) Manual: You may set up the paths manually. First, open MATLAB and go to File -> Set Path. Click 'Add folder' and select the following folders.
(COMKATROOT is the root folder that you unzip COMKAT files under.)
After that, click 'Add subfolders' and select COMKATROOT/ImageReadingFunctions.
4. Re-start MATLAB. Go to "File -> Set Path". Check if the COMKAT folders are in the paths. If not, that means you do not have the 'owner' access to a file called "pathdef.m" on your system. To solve this, type 'which pathdef.m' in MATLAB. It will show the location of this file. Ask your system administrator to give you write access to this file, and then set the path again.
5. In MATLAB, execute command 'validate'. Validation suite should pop up and allow you to examine if results are correctly generated on your computer.
6. If you are having trouble running the validation suite associated with'pxEval' or 'cvodesGenmod5' functions, that means you have a compatibility issue with these mex function. It's either that your MATLAB version is older than the one we used to compile these mex functions, or your platform is not supported by COMKAT yet. You may use the makefile to re-compile these mex functions again. Run the script 'makefile.m' under COMKAT root folder. Please contact us if you are having trouble running it.
NOTE: If you are running MATLAB versions before R2006, we strongly recommend you to re-compile the mex functions again to ensure every function is working properly.
Within the MATLAB File -> Set Path tool you might need to use "Move Up" and "Move Down" to get the folders into the preferred order given above.
4. In MATLAB, execute command 'validate'. Validation Suite should pop up and allow you to select various computations to perform. The generated results are shown in a report displayed in a Web browser. Pay particular attention to the Results section of the report and compare new results generated on Your Computer (to the left) to ones previously generated on a COMKAT Developer's Computer (to the right) and deemed to be correct. Note that minor differences are expected due to finite numeric precision obtained with computers.
NOTE 64-bit Windows users: When you install 64-bit MATLAB, it does not automatically install the Visual C re-distributable for you. But for 32-bit MATLAB it does. Since we compiled our mex files using Visual Studio 2008 for Windows, you will have to install the Visual Studio 2008 re-distributable in order to run the mex files. Otherwise you may get a mex file error.
Please go to Microsoft website: http://www.microsoft.com/downloads/details.aspx?familyid=bd2a6171-e2d6-4230-b809-9a8d7548c1b6&displaylang=en to download and install the re-distributables. If the link is broken, google Microsoft Visual C++ 2008 Redistributable Package x64.