~This is the sample implementation of a Markov random field (MRF) basedimage segmentation algorithm. The main code (mrf.cpp) has been writtenby Csaba Gradwohl (Gradwohl.Csaba@stud.u-szeged.hu) with some minorcontributions from Zoltan Kato (kato@inf.u-szeged.hu,http://www.inf.u-szeged.hu/~kato/). This code is released under theGNU General Public License (seehttp://www.gnu.org/copyleft/gpl.html). Please acknowledge the use ofour program by refering to the following papers (they contain detailedinformation about the MRF model and optimization algorithms):1) M. Berthod, Z. Kato, S. Yu, J. Zerubia: Bayesian imageclassification using Markov random fields. Image and Vision Computing,14(1996): 285-295, 1996.2) Z. Kato: Multi-scale Markovian Modelisation in Computer Vision withApplications to SPOT Image Segmentation. PhD thesis, INRIA SophiaAntipolis, France, 1994.3) Z. Kato, J. Zerubia and M. Berthod: Satellite image classification using amodified Metropolis dynamics Proc. IEEE International Conf. on Acoust., Speechand Sig. Proc., vol. 3, pp. 573-576, San Francisco, CA, March 23-26,1992.The program uses the "Mersenne Twister" random number generatorwritten by Agner Fog (http://www.agner.org/random/). The generatoritself is described in the article by M. Matsumoto & T. Nishimura, in:ACM Transactions on Modeling and Computer Simulation, vol. 8, no. 1,1998, pp. 3-30. Details on the initialization scheme can be found athttp://www.math.sci.hiroshima-u.ac.jp/~m-mat/MT/emt.html .INSTALLATION:=============The code is platform independent. We have succesfully compiled itunder Linux (RedHat, Fedora, Debian) and Windows (XP, 2003 server).1) You need wxWindows 2.4.2 or later (http://www.wxwindows.org)2a) Under Linux/Unix, please edit the makefile and then	$ make clean	$ make   should compile and install the program (it is called mrfdemo).2b) Under Windows, we provide both VC++6 as well as .NET (VC++7)    compatible project files. Open the right "mrfdemo" workspace    (resp. solution) file and choose "Batch Build..." from the menu to    compile the program. Make sure the environment variable "WXWIN" is set    correctly, otherwise you have to modife the project settings.USAGE NOTES:============The program works on BMP images. Some test images are provided under the'images' subdirectory. The program GUI should be intuitive. Mainsteps:1) Load an image2) Enter the number of pixel classes (~region type)3) Push "Select classes" button4) Press left mouse button over the input image and draw a rectangleover a representative region of the first class. Then push "Nextclass" button. The mean and variance should appear in the "Classparameters" window. Continue with the next class until arepresentative rectangle for all classes has been selected.5) Set the weight of doubleton potentials (default is 0.9) and thestopping threshold (iterations are stopped when the energy change isless than the specified value).6) Choose the optimization method from the pull-down list.7) Adjust the optimization method's parameters:	T0    - Initial temperature	c     - temperature scheduler (T(n+1) = c*T(n))	alpha - MMD's probability threshold8) Push "Do it >>" button to execute segmentation.9) Optionally, you can save the segmentation result as a BMP image.During segmentation, the current classification along with thetemperature and global energy are displayed at each iteration. At theend, the elapsed CPU time is also displayed (excluding GUI oveheaad!).