<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html><head><meta http-equiv="Content-Type" content="text/html;charset=iso-8859-1">
<title>model_search_impl.h Source File</title>
<link href="doxygen.css" rel="stylesheet" type="text/css">
<link href="tabs.css" rel="stylesheet" type="text/css">
</head><body>
<div align="left"><a href="http://www.mrpt.org/">Main MRPT website</a> > <b>C++ reference</b> </div>
<div align="right">
<a href="index.html"><img border="0" src="mrpt_logo.png" alt="MRPT logo"></a>
</div>
<!-- Generated by Doxygen 1.7.5 -->
<script type="text/javascript">
var searchBox = new SearchBox("searchBox", "search",false,'Search');
</script>
<div id="navrow1" class="tabs">
<ul class="tablist">
<li><a href="index.html"><span>Main Page</span></a></li>
<li><a href="pages.html"><span>Related Pages</span></a></li>
<li><a href="modules.html"><span>Modules</span></a></li>
<li><a href="namespaces.html"><span>Namespaces</span></a></li>
<li><a href="annotated.html"><span>Classes</span></a></li>
<li class="current"><a href="files.html"><span>Files</span></a></li>
<li>
<div id="MSearchBox" class="MSearchBoxInactive">
<div class="left">
<form id="FSearchBox" action="search.php" method="get">
<img id="MSearchSelect" src="search/mag.png" alt=""/>
<input type="text" id="MSearchField" name="query" value="Search" size="20" accesskey="S"
onfocus="searchBox.OnSearchFieldFocus(true)"
onblur="searchBox.OnSearchFieldFocus(false)"/>
</form>
</div><div class="right"></div>
</div>
</li>
</ul>
</div>
<div id="navrow2" class="tabs2">
<ul class="tablist">
<li><a href="files.html"><span>File List</span></a></li>
<li><a href="globals.html"><span>File Members</span></a></li>
</ul>
</div>
<div class="header">
<div class="headertitle">
<div class="title">model_search_impl.h</div> </div>
</div>
<div class="contents">
<a href="model__search__impl_8h.html">Go to the documentation of this file.</a><div class="fragment"><pre class="fragment"><a name="l00001"></a>00001 <span class="comment">/* +---------------------------------------------------------------------------+</span>
<a name="l00002"></a>00002 <span class="comment"> | The Mobile Robot Programming Toolkit (MRPT) C++ library |</span>
<a name="l00003"></a>00003 <span class="comment"> | |</span>
<a name="l00004"></a>00004 <span class="comment"> | http://www.mrpt.org/ |</span>
<a name="l00005"></a>00005 <span class="comment"> | |</span>
<a name="l00006"></a>00006 <span class="comment"> | Copyright (C) 2005-2011 University of Malaga |</span>
<a name="l00007"></a>00007 <span class="comment"> | |</span>
<a name="l00008"></a>00008 <span class="comment"> | This software was written by the Machine Perception and Intelligent |</span>
<a name="l00009"></a>00009 <span class="comment"> | Robotics Lab, University of Malaga (Spain). |</span>
<a name="l00010"></a>00010 <span class="comment"> | Contact: Jose-Luis Blanco <jlblanco@ctima.uma.es> |</span>
<a name="l00011"></a>00011 <span class="comment"> | |</span>
<a name="l00012"></a>00012 <span class="comment"> | This file is part of the MRPT project. |</span>
<a name="l00013"></a>00013 <span class="comment"> | |</span>
<a name="l00014"></a>00014 <span class="comment"> | MRPT is free software: you can redistribute it and/or modify |</span>
<a name="l00015"></a>00015 <span class="comment"> | it under the terms of the GNU General Public License as published by |</span>
<a name="l00016"></a>00016 <span class="comment"> | the Free Software Foundation, either version 3 of the License, or |</span>
<a name="l00017"></a>00017 <span class="comment"> | (at your option) any later version. |</span>
<a name="l00018"></a>00018 <span class="comment"> | |</span>
<a name="l00019"></a>00019 <span class="comment"> | MRPT is distributed in the hope that it will be useful, |</span>
<a name="l00020"></a>00020 <span class="comment"> | but WITHOUT ANY WARRANTY; without even the implied warranty of |</span>
<a name="l00021"></a>00021 <span class="comment"> | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |</span>
<a name="l00022"></a>00022 <span class="comment"> | GNU General Public License for more details. |</span>
<a name="l00023"></a>00023 <span class="comment"> | |</span>
<a name="l00024"></a>00024 <span class="comment"> | You should have received a copy of the GNU General Public License |</span>
<a name="l00025"></a>00025 <span class="comment"> | along with MRPT. If not, see <http://www.gnu.org/licenses/>. |</span>
<a name="l00026"></a>00026 <span class="comment"> | |</span>
<a name="l00027"></a>00027 <span class="comment"> +---------------------------------------------------------------------------+ */</span>
<a name="l00028"></a>00028
<a name="l00029"></a>00029 <span class="preprocessor">#ifndef math_modelsearch_impl_h</span>
<a name="l00030"></a>00030 <span class="preprocessor"></span><span class="preprocessor">#define math_modelsearch_impl_h</span>
<a name="l00031"></a>00031 <span class="preprocessor"></span>
<a name="l00032"></a>00032 <span class="preprocessor">#ifndef math_modelsearch_h</span>
<a name="l00033"></a>00033 <span class="preprocessor"></span><span class="preprocessor"># include "<a class="code" href="model__search_8h.html">model_search.h</a>"</span>
<a name="l00034"></a>00034 <span class="preprocessor">#endif</span>
<a name="l00035"></a>00035 <span class="preprocessor"></span>
<a name="l00036"></a>00036 <span class="preprocessor">#include <limits></span>
<a name="l00037"></a>00037
<a name="l00038"></a>00038 <span class="keyword">namespace </span>mrpt {
<a name="l00039"></a>00039 <span class="keyword">namespace </span>math {
<a name="l00040"></a>00040
<a name="l00041"></a>00041 <span class="comment">//----------------------------------------------------------------------</span><span class="comment"></span>
<a name="l00042"></a>00042 <span class="comment">//! Run the ransac algorithm searching for a single model</span>
<a name="l00043"></a>00043 <span class="comment"></span><span class="keyword">template</span><<span class="keyword">typename</span> TModelFit>
<a name="l00044"></a><a class="code" href="classmrpt_1_1math_1_1_model_search.html#aa4de9616da8e038ffa32a92959780488">00044</a> <span class="keywordtype">bool</span> <a class="code" href="classmrpt_1_1math_1_1_model_search.html#aa4de9616da8e038ffa32a92959780488" title="Run the ransac algorithm searching for a single model.">ModelSearch::ransacSingleModel</a>( <span class="keyword">const</span> TModelFit& p_state,
<a name="l00045"></a>00045 <span class="keywordtype">size_t</span> p_kernelSize,
<a name="l00046"></a>00046 <span class="keyword">const</span> <span class="keyword">typename</span> TModelFit::Real& p_fitnessThreshold,
<a name="l00047"></a>00047 <span class="keyword">typename</span> TModelFit::Model& p_bestModel,
<a name="l00048"></a>00048 <a class="code" href="classstd_1_1vector.html">vector_size_t</a>& p_inliers )
<a name="l00049"></a>00049 {
<a name="l00050"></a>00050 <span class="keywordtype">size_t</span> bestScore = 0;
<a name="l00051"></a>00051 <span class="keywordtype">size_t</span> iter = 0;
<a name="l00052"></a>00052 <span class="keywordtype">size_t</span> softIterLimit = 1; <span class="comment">// will be updated by the size of inliers</span>
<a name="l00053"></a>00053 <span class="keywordtype">size_t</span> hardIterLimit = 100; <span class="comment">// a fixed iteration step</span>
<a name="l00054"></a>00054 <span class="keywordtype">size_t</span> nSamples = p_state.getSampleCount();
<a name="l00055"></a>00055 <a class="code" href="classstd_1_1vector.html">vector_size_t</a> ind( p_kernelSize );
<a name="l00056"></a>00056
<a name="l00057"></a>00057 <span class="keywordflow">while</span> ( iter < softIterLimit && iter < hardIterLimit )
<a name="l00058"></a>00058 {
<a name="l00059"></a>00059 <span class="keywordtype">bool</span> degenerate = <span class="keyword">true</span>;
<a name="l00060"></a>00060 <span class="keyword">typename</span> TModelFit::Model currentModel;
<a name="l00061"></a>00061 <span class="keywordtype">size_t</span> i = 0;
<a name="l00062"></a>00062 <span class="keywordflow">while</span> ( degenerate )
<a name="l00063"></a>00063 {
<a name="l00064"></a>00064 <a class="code" href="classmrpt_1_1math_1_1_model_search.html#a0fecc75e49307a584d0f2e21ce679414" title="Select random (unique) indices from the 0..p_size sequence.">pickRandomIndex</a>( nSamples, p_kernelSize, ind );
<a name="l00065"></a>00065 degenerate = !p_state.fitModel( ind, currentModel );
<a name="l00066"></a>00066 i++;
<a name="l00067"></a>00067 <span class="keywordflow">if</span>( i > 100 )
<a name="l00068"></a>00068 <span class="keywordflow">return</span> <span class="keyword">false</span>;
<a name="l00069"></a>00069 }
<a name="l00070"></a>00070
<a name="l00071"></a>00071 <a class="code" href="classstd_1_1vector.html">vector_size_t</a> inliers;
<a name="l00072"></a>00072
<a name="l00073"></a>00073 <span class="keywordflow">for</span>( <span class="keywordtype">size_t</span> i = 0; i < nSamples; i++ )
<a name="l00074"></a>00074 {
<a name="l00075"></a>00075 <span class="keywordflow">if</span>( p_state.testSample( i, currentModel ) < p_fitnessThreshold )
<a name="l00076"></a>00076 inliers.push_back( i );
<a name="l00077"></a>00077 }
<a name="l00078"></a>00078 <a class="code" href="mrpt__macros_8h.html#a47eb5a445c2bf3d9190396510ea9683e">ASSERT_</a>( inliers.size() > 0 );
<a name="l00079"></a>00079
<a name="l00080"></a>00080 <span class="comment">// Find the number of inliers to this model.</span>
<a name="l00081"></a>00081 <span class="keyword">const</span> <span class="keywordtype">size_t</span> ninliers = inliers.size();
<a name="l00082"></a>00082
<a name="l00083"></a>00083 <span class="keywordflow">if</span> ( ninliers > bestScore )
<a name="l00084"></a>00084 {
<a name="l00085"></a>00085 bestScore = ninliers;
<a name="l00086"></a>00086 p_bestModel = currentModel;
<a name="l00087"></a>00087 p_inliers = inliers;
<a name="l00088"></a>00088
<a name="l00089"></a>00089 <span class="comment">// Update the estimation of maxIter to pick dataset with no outliers at propability p</span>
<a name="l00090"></a>00090 <span class="keywordtype">float</span> f = ninliers / <span class="keyword">static_cast<</span><span class="keywordtype">float</span><span class="keyword">></span>( nSamples );
<a name="l00091"></a>00091 <span class="keywordtype">float</span> p = 1 - pow( f, static_cast<float>( p_kernelSize ) );
<a name="l00092"></a>00092 <span class="keywordtype">float</span> eps = std::numeric_limits<float>::epsilon();
<a name="l00093"></a>00093 p = std::max( eps, p); <span class="comment">// Avoid division by -Inf</span>
<a name="l00094"></a>00094 p = std::min( 1-eps, p); <span class="comment">// Avoid division by 0.</span>
<a name="l00095"></a>00095 softIterLimit = log(1-p) / log(p);
<a name="l00096"></a>00096 }
<a name="l00097"></a>00097
<a name="l00098"></a>00098 iter++;
<a name="l00099"></a>00099 }
<a name="l00100"></a>00100
<a name="l00101"></a>00101 <span class="keywordflow">return</span> <span class="keyword">true</span>;
<a name="l00102"></a>00102 }
<a name="l00103"></a>00103
<a name="l00104"></a>00104 <span class="comment">//----------------------------------------------------------------------</span><span class="comment"></span>
<a name="l00105"></a>00105 <span class="comment">//! Run a generic programming version of ransac searching for a single model</span>
<a name="l00106"></a>00106 <span class="comment"></span><span class="keyword">template</span><<span class="keyword">typename</span> TModelFit>
<a name="l00107"></a><a class="code" href="classmrpt_1_1math_1_1_model_search.html#a56cf13ab97fef4a0e57660fb7cbd54c4">00107</a> <span class="keywordtype">bool</span> <a class="code" href="classmrpt_1_1math_1_1_model_search.html#a56cf13ab97fef4a0e57660fb7cbd54c4" title="Run a generic programming version of ransac searching for a single model.">ModelSearch::geneticSingleModel</a>( <span class="keyword">const</span> TModelFit& p_state,
<a name="l00108"></a>00108 <span class="keywordtype">size_t</span> p_kernelSize,
<a name="l00109"></a>00109 <span class="keyword">const</span> <span class="keyword">typename</span> TModelFit::Real& p_fitnessThreshold,
<a name="l00110"></a>00110 <span class="keywordtype">size_t</span> p_populationSize,
<a name="l00111"></a>00111 <span class="keywordtype">size_t</span> p_maxIteration,
<a name="l00112"></a>00112 <span class="keyword">typename</span> TModelFit::Model& p_bestModel,
<a name="l00113"></a>00113 <a class="code" href="classstd_1_1vector.html">vector_size_t</a>& p_inliers )
<a name="l00114"></a>00114 {
<a name="l00115"></a>00115 <span class="comment">// a single specie of the population</span>
<a name="l00116"></a>00116 <span class="keyword">typedef</span> <a class="code" href="structmrpt_1_1math_1_1_model_search_1_1_t_species.html">TSpecies<TModelFit></a> Species;
<a name="l00117"></a>00117 std<a class="code" href="classstd_1_1vector.html" title="STL class.">::vector<Species></a> storage;
<a name="l00118"></a>00118 std<a class="code" href="classstd_1_1vector.html" title="STL class.">::vector<Species*></a> population;
<a name="l00119"></a>00119 std<a class="code" href="classstd_1_1vector.html" title="STL class.">::vector<Species*></a> sortedPopulation;
<a name="l00120"></a>00120
<a name="l00121"></a>00121 <span class="keywordtype">size_t</span> sampleCount = p_state.getSampleCount();
<a name="l00122"></a>00122 <span class="keywordtype">int</span> elderCnt = (int)p_populationSize/3;
<a name="l00123"></a>00123 <span class="keywordtype">int</span> siblingCnt = (p_populationSize - elderCnt) / 2;
<a name="l00124"></a>00124 <span class="keywordtype">int</span> speciesAlive = 0;
<a name="l00125"></a>00125
<a name="l00126"></a>00126 storage.resize( p_populationSize );
<a name="l00127"></a>00127 population.reserve( p_populationSize );
<a name="l00128"></a>00128 sortedPopulation.reserve( p_populationSize );
<a name="l00129"></a>00129 <span class="keywordflow">for</span>( <span class="keyword">typename</span> <a class="code" href="classstd_1_1vector_1_1iterator.html" title="STL iterator class.">std::vector<Species>::iterator</a> it = storage.begin(); it != storage.end(); it++ )
<a name="l00130"></a>00130 {
<a name="l00131"></a>00131 <a class="code" href="classmrpt_1_1math_1_1_model_search.html#a0fecc75e49307a584d0f2e21ce679414" title="Select random (unique) indices from the 0..p_size sequence.">pickRandomIndex</a>( sampleCount, p_kernelSize, it->sample );
<a name="l00132"></a>00132 population.push_back( &*it );
<a name="l00133"></a>00133 sortedPopulation.push_back( &*it );
<a name="l00134"></a>00134 }
<a name="l00135"></a>00135
<a name="l00136"></a>00136 <span class="keywordtype">size_t</span> iter = 0;
<a name="l00137"></a>00137 <span class="keywordflow">while</span> ( iter < p_maxIteration )
<a name="l00138"></a>00138 {
<a name="l00139"></a>00139 <span class="keywordflow">if</span>( iter > 0 )
<a name="l00140"></a>00140 {
<a name="l00141"></a>00141 <span class="comment">//generate new population: old, siblings, new</span>
<a name="l00142"></a>00142 population.clear();
<a name="l00143"></a>00143 <span class="keywordtype">int</span> i = 0;
<a name="l00144"></a>00144
<a name="l00145"></a>00145 <span class="comment">//copy the best elders</span>
<a name="l00146"></a>00146 <span class="keywordflow">for</span>(; i < elderCnt; i++ )
<a name="l00147"></a>00147 {
<a name="l00148"></a>00148 population.push_back( sortedPopulation[i] );
<a name="l00149"></a>00149 }
<a name="l00150"></a>00150
<a name="l00151"></a>00151 <span class="comment">// mate elders to make siblings</span>
<a name="l00152"></a>00152 <span class="keywordtype">int</span> se = (int)speciesAlive; <span class="comment">//dead species cannot mate</span>
<a name="l00153"></a>00153 <span class="keywordflow">for</span>( ; i < elderCnt + siblingCnt ; i++ )
<a name="l00154"></a>00154 {
<a name="l00155"></a>00155 Species* sibling = sortedPopulation[--se];
<a name="l00156"></a>00156 population.push_back( sibling );
<a name="l00157"></a>00157
<a name="l00158"></a>00158 <span class="comment">//pick two parents, from the species not yet refactored</span>
<a name="l00159"></a>00159 <span class="comment">//better elders has more chance to mate as they are removed later from the list</span>
<a name="l00160"></a>00160 <span class="keywordtype">int</span> r1 = rand();
<a name="l00161"></a>00161 <span class="keywordtype">int</span> r2 = rand();
<a name="l00162"></a>00162 <span class="keywordtype">int</span> p1 = r1 % se;
<a name="l00163"></a>00163 <span class="keywordtype">int</span> p2 = (p1 > se / 2) ? (r2 % p1) : p1 + 1 + (r2 % (se-p1-1));
<a name="l00164"></a>00164 <a class="code" href="mrpt__macros_8h.html#a47eb5a445c2bf3d9190396510ea9683e">ASSERT_</a>( p1 != p2 && p1 < se && p2 < se );
<a name="l00165"></a>00165 <a class="code" href="mrpt__macros_8h.html#a47eb5a445c2bf3d9190396510ea9683e">ASSERT_</a>( se >= elderCnt );
<a name="l00166"></a>00166 Species* a = sortedPopulation[p1];
<a name="l00167"></a>00167 Species* b = sortedPopulation[p2];
<a name="l00168"></a>00168
<a name="l00169"></a>00169 <span class="comment">// merge the sample candidates</span>
<a name="l00170"></a>00170 std<a class="code" href="classstd_1_1set.html" title="STL class.">::set<size_t></a> sampleSet;
<a name="l00171"></a>00171 sampleSet.insert( a->sample.begin(), a->sample.end() );
<a name="l00172"></a>00172 sampleSet.insert( b->sample.begin(), b->sample.end() );
<a name="l00173"></a>00173 <span class="comment">//mutate - add a random sample that will be selected with some (non-zero) probability</span>
<a name="l00174"></a>00174 sampleSet.insert( rand() % sampleCount );
<a name="l00175"></a>00175 <a class="code" href="classmrpt_1_1math_1_1_model_search.html#a0fecc75e49307a584d0f2e21ce679414" title="Select random (unique) indices from the 0..p_size sequence.">pickRandomIndex</a>( sampleSet, p_kernelSize, sibling->sample );
<a name="l00176"></a>00176 }
<a name="l00177"></a>00177
<a name="l00178"></a>00178 <span class="comment">// generate some new random species</span>
<a name="l00179"></a>00179 <span class="keywordflow">for</span>( ; i < (int)p_populationSize; i++ )
<a name="l00180"></a>00180 {
<a name="l00181"></a>00181 Species* s = sortedPopulation[i];
<a name="l00182"></a>00182 population.push_back( s );
<a name="l00183"></a>00183 <a class="code" href="classmrpt_1_1math_1_1_model_search.html#a0fecc75e49307a584d0f2e21ce679414" title="Select random (unique) indices from the 0..p_size sequence.">pickRandomIndex</a>( sampleCount, p_kernelSize, s->sample );
<a name="l00184"></a>00184 }
<a name="l00185"></a>00185 }
<a name="l00186"></a>00186
<a name="l00187"></a>00187 <span class="comment">//evaluate species</span>
<a name="l00188"></a>00188 speciesAlive = 0;
<a name="l00189"></a>00189 <span class="keywordflow">for</span>( <span class="keyword">typename</span> <a class="code" href="classstd_1_1vector_1_1iterator.html" title="STL iterator class.">std::vector<Species*>::iterator</a> it = population.begin(); it != population.end(); it++ )
<a name="l00190"></a>00190 {
<a name="l00191"></a>00191 Species& s = **it;
<a name="l00192"></a>00192 <span class="keywordflow">if</span>( p_state.fitModel( s.sample, s.model ) )
<a name="l00193"></a>00193 {
<a name="l00194"></a>00194 s.fitness = 0;
<a name="l00195"></a>00195 <span class="keywordflow">for</span>( <span class="keywordtype">size_t</span> i = 0; i < p_state.getSampleCount(); i++ )
<a name="l00196"></a>00196 {
<a name="l00197"></a>00197 <span class="keyword">typename</span> TModelFit::Real f = p_state.testSample( i, s.model );
<a name="l00198"></a>00198 <span class="keywordflow">if</span>( f < p_fitnessThreshold )
<a name="l00199"></a>00199 {
<a name="l00200"></a>00200 s.fitness += f;
<a name="l00201"></a>00201 s.inliers.push_back( i );
<a name="l00202"></a>00202 }
<a name="l00203"></a>00203 }
<a name="l00204"></a>00204 <a class="code" href="mrpt__macros_8h.html#a47eb5a445c2bf3d9190396510ea9683e">ASSERT_</a>( s.inliers.size() > 0 );
<a name="l00205"></a>00205
<a name="l00206"></a>00206 s.fitness /= s.inliers.size();
<a name="l00207"></a>00207 <span class="comment">// scale by the number of outliers</span>
<a name="l00208"></a>00208 s.fitness *= (sampleCount - s.inliers.size());
<a name="l00209"></a>00209 speciesAlive++;
<a name="l00210"></a>00210 }
<a name="l00211"></a>00211 <span class="keywordflow">else</span>
<a name="l00212"></a>00212 s.fitness = std::numeric_limits<typename TModelFit::Real>::max();
<a name="l00213"></a>00213 }
<a name="l00214"></a>00214
<a name="l00215"></a>00215 <span class="keywordflow">if</span>( !speciesAlive )
<a name="l00216"></a>00216 {
<a name="l00217"></a>00217 <span class="comment">// the world is dead, no model was found</span>
<a name="l00218"></a>00218 <span class="keywordflow">return</span> <span class="keyword">false</span>;
<a name="l00219"></a>00219 }
<a name="l00220"></a>00220
<a name="l00221"></a>00221 <span class="comment">//sort by fitness (ascending)</span>
<a name="l00222"></a>00222 std::sort( sortedPopulation.begin(), sortedPopulation.end(), Species::compare );
<a name="l00223"></a>00223
<a name="l00224"></a>00224 iter++;
<a name="l00225"></a>00225 }
<a name="l00226"></a>00226
<a name="l00227"></a>00227 p_bestModel = sortedPopulation[0]->model;
<a name="l00228"></a>00228 p_inliers = sortedPopulation[0]->inliers;
<a name="l00229"></a>00229
<a name="l00230"></a>00230 <span class="keywordflow">return</span> !p_inliers.empty();
<a name="l00231"></a>00231 }
<a name="l00232"></a>00232
<a name="l00233"></a>00233 } <span class="comment">// namespace math</span>
<a name="l00234"></a>00234 } <span class="comment">// namespace mrpt</span>
<a name="l00235"></a>00235
<a name="l00236"></a>00236 <span class="preprocessor">#endif // math_modelsearch_h</span>
</pre></div></div>
</div>
<br><hr><br> <table border="0" width="100%"> <tr> <td> Page generated by <a href="http://www.doxygen.org" target="_blank">Doxygen 1.7.5</a> for MRPT 0.9.5 SVN: at Sun Sep 25 17:20:18 UTC 2011</td><td></td> <td width="100"> </td> <td width="150"> </td></tr> </table> </body></html>
