readme.txt

快速高斯变换的程序

Improved Fast Gauss Transform, April 2004, Version 1.01

Author: Changjiang Yang

Department of Computer Science
University of Maryland
College Park, MD 20742
USA

Contact: [e-mail], [e-mail], [e-mail]

--------------------------------------------------------------------------

URL: http://www.cs.umd.edu/~yangcj/node1.html

+-------------------------------------------------------------------------+
|			       SOURCE CODE			          |
+-------------------------------------------------------------------------+

The source code is written with both Matlab and C++. The Matlab code is for the interface, and the C++ code implements the Improved Fast Gauss Transform.

MATLAB Source Files:

Test4.m		Test program for image segmentation with mean-shift algorithm
Rgb2luv.m	Convert RGB image to Lu*v* image
Meanshift.m	Mean-shift algorithm speeded up by improved fast Gauss transform

C++ Source Files:

Kctfgt.cpp 		The C Mex-file for MATLAB
FastGauss.h		The header file for FastGauss.cpp
FastGauss.cpp		The C++ implementation for improved fast Gauss transform
Kcenter.cpp		The C++ implementation for k-center algorithm

Smhouse.ppm		The test image

+-------------------------------------------------------------------------+
|			COMPILATION			    	          |
+-------------------------------------------------------------------------+

You need to compile the C++ source files into mex-file. 

1.Compiler Requirements:
We tested our programs on Windows, Solaris and Linux.
Platform		Compiler		Mex-File Extension
Linux			gcc			mexglx
Solaris			gcc			mexsol
Windows			Visual C++ 6.0		dll

2.Building Mex-File:
Under Solaris or Linux, at the MATLAB prompt, type
	Mex kctfgt.cpp FastGauss.cpp kcenter.cpp
Under Windows, use Microsoft Visual C++ 6.0 to open the kctfgt.dsw, go to 揃uild->Build kctfgt.dll