Pattern Recognition Techniques
2013/2014
Practical Exercises : Class # 1 and #2
Statistical Pattern Recognition Toolbox - (STPRTool): Installation & Data
Handling
STPRTool Requirements
• Matlab, version 5.3 or higher (Verify your Matlab version by typing in the
Matlab command line ver );
• Some algorithms require the Optimization and Image Processing Toolboxes (Verify if you have it installed by typing the command ver).
STPRTool Installation
1. Download STPRTool Version 2.12 (2-sep-2013) at http://cmp.felk.cvut.cz/cmp/software/stprtool/stprtool12sep13.zip; 2. Create a directory and decompress the downloaded file. It is recommended to create a folder in the [Matlab Root]/toolbox directory;
3. Set the path to the toolbox by running the stprpath.m script. Here you have two options:
• Set the path manually each time that you initiate Matlab: change the Matlab Current Folder to the folder were STPRTool is located and run stprpath in the command line;
• Do it automatically by using the startup.m file. The startup.m file exists or should be created in the Matlab Startup Folder and executes automatically a list of instructions when Matlab starts. Add the following instructions to the startup.m file:
– eval([’cd ’,fullfile(matlabroot,’toolbox’,’stprtool’)]);%Change current directory to the toolbox folder
– stprpath %Add STPRTool paths
– upath=userpath;%Store the default startup path
– eval([’cd ’,upath(1:end-1)]);%Return back to the default folder
4. Compile C language algorithms. To do this change the current directory to the STPRTool directory and run compilemex in the command line.
1
5. Now you should have all the STPRTool functionalities available. Type help stprtool to see the list of functionalities.
Data handling
Generic Structure
During this course we will follow the same data storing structure as used by the
STPRTool. This structure contains by default 5 fields:
• .X: Matrix