
SMAP version 1.6.1
	Contains:
	1) Conversion to full ANSI C
	2) Support of pgm file format for images
	3) An improved version of the cluster program
	4) A new example (cluster-3.1/example2) that shows how the
	cluster and and smap programs may be used together to segment color images.
SMAP version 1.6.2: Includes update to cluster-3.5.4
SMAP version 1.6.3: Includes reorganized/improved examples
SMAP version 1.6.4: Fixes bug in extract subroutine


This directory contains three subdirectories 
documentation/ - contains a user_manual which explains how to use this software

cluster-X.X.X/ - a clustering algorithm used to generate class model parameters
              from training data. This program is also generally useful
              as an unsupervised clustering program.

segment/ - the SMAP segmentation algorithm documented in the jounal publication:
           C. Bouman, M. Shapiro
           "A Multiscale Model for Bayesian Image Segmentation"
           IEEE Trans. on Image Processing
           vol. 3, no. 2, pp 162-177, March 1994.

pgm/ - This subdirectory contains ANSII C subroutines for reading and
       writing pgm, ppm, and pbm file formats


The SMAP algorithm may be used to segment single band or multispectral
image data. Each image class is specified as a Gaussian-mixture model.
The parameters of the Gaussian-mixture model may be specified manually
(see the example_param_file) or automatically from training data using
the clust program in the cluster subdirectory. 

You can run an example by compiling the code 
in the cluster-X.X.X/src and segment/src subdirectories
and running the example shell scripts in cluster/exampleX and segment/exampleX.
You can also reproduce the results from the March 94 paper by running
the shell script in segment/example2




For more information contact:
  Charles A. Bouman
  School of Electrical Engineering
  Purdue University
  West Lafayette IN 47906
  bouman@ecn.purdue.edu
  http://www.ece.purdue.edu/~bouman/
