medical image analysis SVN (old)

/* -*- mona-c++  -*-
 * Copyright (c) Leipzig, Madrid 2004 - 2008 
 * Max-Planck-Institute for Human Cognitive and Brain Science	
 * Max-Planck-Institute for Evolutionary Anthropology 
 * BIT, ETSI Telecomunicacion, UPM
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 *
 */

#include <libmona/monaIOStream.hh>
#include <map>
#include <stdexcept>
#include <cmath>


typedef std::pair<int, int> CClassRange; 
typedef std::map<double, CClassRange> CClassMap; 

CClassMap kmeans(size_t classes, const std::vector<double>& histo, int max_iter, bool even_start = true)
{
	/** now kmeans */
	std::vector<double> center(classes);

	if (even_start) {
		double steps = histo.size() / double(classes); 
		for (size_t i = 0; i < classes; ++i)
			center[i] = (i+1) * steps; 
	}else{
		
		size_t intensity = 0; 
		double g_size = 1.0  / classes;
		
		size_t i; 
		for (i = 0; (i < classes - 1) && (intensity < histo.size()); ++i) {
			double val = 0.0; 
			double l_size = 0; 
			
			while (l_size < g_size && intensity < histo.size()) {
				l_size += histo[intensity];
				val += histo[intensity] * intensity;
			++intensity;
			}
			center[i] = val / l_size; 
			mona::cvmsg() << "center[" << i << "] = " << center[i] << "\n"; 
		}
		{	
			double val = 0.0; 
			double l_size = 0; 
			
			while (intensity < histo.size()) {
				l_size += histo[intensity];
				val += histo[intensity] * intensity;
				++intensity;
				
			}
			center[i] = val / l_size; 
			mona::cvmsg() << "center[" << i << "] = " << center[i] << "\n"; 
		}
	}
		
	std::vector<size_t> next_center(histo.size());

	double diff = 10.0;
	int iter = 0; 

	while (diff > 1.0 && iter++ < max_iter) {
		mona::cvdebug() << "iterate classes:" << iter << "\n"; 
		diff = 0.0; 
		
		for (size_t i = 0; i < next_center.size(); ++i) {
			next_center[i] = 0; 
			double dist = fabs(i - center[0]);
			for (size_t k = 1; k < classes; ++k) {
				double ndist = fabs(i - center[k]);
				if (ndist < dist) {
					dist = ndist;
					next_center[i] = k; 
				}
			}
		}
		
		for (size_t i = 0; i < classes; ++i) {
			double xnorm  = 0.0; 
			double sum = 0.0; 

			for (size_t j = 0; j < next_center.size(); j++)
				if (next_center[j] == i) {
					xnorm += histo[j];
					sum += j * histo[j];
				}
			
			if (xnorm > 0)
				sum /= xnorm;
			else
				sum = 0;
			
			double dx = sum - center[i];
			center[i] = sum;
			diff += dx * dx;
			mona::cvdebug() << "center[" << i << "] = " << center[i] << "\n"; 
		}
		
		mona::cvdebug() << "diff = " << diff << std::endl;  
	}
	
	if (next_center[0] != 0)
		throw std::runtime_error("First value not next to the first center - this shouldn't happen: Bye bye\n");
  


	CClassMap result; 
	CClassRange range; 
	range.first = range.second = 0; 
	for (size_t i = 0; i < classes; ++i) {
		while (next_center[range.second] == i) 
			++range.second; 
		result[center[i]] = range; 
		range.first = range.second; 
	}
	return result; 
}

int main(int argc, const char *args[])
{
	int classes; 
	int max_iter; 
	bool even_start; 
	
	popt::COptions options;
	options.push_back(popt::option( classes, "classes", 'n', "number of classes to partition into","10")); 
	options.push_back(popt::option( max_iter, "max-iter", 'm', "maximum number of iterations", "100")); 
	options.push_back(popt::option( even_start, "even-start", 'e', "start with centers evenly distributed over the histogram")); 
	

	
	try {
		std::vector<std::string> non_args; 
		
		popt::parse_options(argc, args, options, non_args);
		
		if (!non_args.empty()) {
			mona::cverr() << "some arguments where not recogniced\n"; 
			mona::cverr() << "try 'eva-kmeans --help' for help\n";
			return -1; 
		}
		
		std::vector<double> histo; 
		// read input data
		while (std::cin.good()) {
			float val;
			std::cin >> val; 
			histo.push_back(val); 
		}
		
		CClassMap result = kmeans(classes, histo, max_iter, even_start); 
		
		std::cout << result.size() << "\n"; 
		
		for (CClassMap::const_iterator i = result.begin(); i != result.end(); ++i) {
			std::cout << i->first << " " << i->second.first << " " <<  i->second.second <<"\n"; 
		}
		return 0; 
	}
	
	catch (const std::runtime_error &e){
		std::cerr << args[0] << " runtime: " << e.what() << std::endl;
	}
	catch (const std::invalid_argument &e){
		std::cerr << args[0] << " error: " << e.what() << std::endl;
	}
	catch (const std::exception& e){
		std::cerr << args[0] << " error: " << e.what() << std::endl;
	}
	catch (...){
		std::cerr << args[0] << " unknown exception" << std::endl;
	}
	return EXIT_FAILURE;
}