00001 /*!@file Channels/ForegroundDetectionChannel.C Wrapper around OpenCV implementation 00002 * of "Foreground Object Detection from Videos Containing Complex Background" by Huang, 00003 * et. al. in ACMMM 2003*/ 00004 // //////////////////////////////////////////////////////////////////// // 00005 // The iLab Neuromorphic Vision C++ Toolkit - Copyright (C) 2000-2005 // 00006 // by the University of Southern California (USC) and the iLab at USC. // 00007 // See http://iLab.usc.edu for information about this project. // 00008 // //////////////////////////////////////////////////////////////////// // 00009 // Major portions of the iLab Neuromorphic Vision Toolkit are protected // 00010 // under the U.S. patent ``Computation of Intrinsic Perceptual Saliency // 00011 // in Visual Environments, and Applications'' by Christof Koch and // 00012 // Laurent Itti, California Institute of Technology, 2001 (patent // 00013 // pending; application number 09/912,225 filed July 23, 2001; see // 00014 // http://pair.uspto.gov/cgi-bin/final/home.pl for current status). // 00015 // //////////////////////////////////////////////////////////////////// // 00016 // This file is part of the iLab Neuromorphic Vision C++ Toolkit. // 00017 // // 00018 // The iLab Neuromorphic Vision C++ Toolkit is free software; you can // 00019 // redistribute it and/or modify it under the terms of the GNU General // 00020 // Public License as published by the Free Software Foundation; either // 00021 // version 2 of the License, or (at your option) any later version. // 00022 // // 00023 // The iLab Neuromorphic Vision C++ Toolkit is distributed in the hope // 00024 // that it will be useful, but WITHOUT ANY WARRANTY; without even the // 00025 // implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR // 00026 // PURPOSE. See the GNU General Public License for more details. // 00027 // // 00028 // You should have received a copy of the GNU General Public License // 00029 // along with the iLab Neuromorphic Vision C++ Toolkit; if not, write // 00030 // to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, // 00031 // Boston, MA 02111-1307 USA. // 00032 // //////////////////////////////////////////////////////////////////// // 00033 // 00034 // Primary maintainer for this file: Randolph Voorhies 00035 // $HeadURL: svn://isvn.usc.edu/software/invt/trunk/saliency/src/Channels/ForegroundDetectionChannel.C $ 00036 // $Id: ForegroundDetectionChannel.C 14605 2011-03-15 02:25:06Z dparks $ 00037 // 00038 00039 00040 #include "Channels/ForegroundDetectionChannel.H" 00041 00042 #include "Image/DrawOps.H" 00043 #include "Image/Kernels.H" 00044 #include "Image/MathOps.H" 00045 #include "Image/ShapeOps.H" 00046 #include "Channels/ChannelOpts.H" 00047 #include "Component/ModelOptionDef.H" 00048 00049 00050 ForegroundDetectionChannel::ForegroundDetectionChannel(OptionManager& mgr) : 00051 SingleChannel(mgr, "ForegroundDetectionChannel", "ForegroundDetectionChannel", FOREGROUND, rutz::shared_ptr<PyrBuilder<float> >()), 00052 itsMap(), 00053 itsLevelSpec(&OPT_LevelSpec, this) 00054 { 00055 #ifdef HAVE_OPENCV 00056 itsStatModel_cv = NULL; 00057 #else 00058 LFATAL("OpenCV is needed for Foreground Detection Channel!"); 00059 #endif 00060 } 00061 00062 // ###################################################################### 00063 ForegroundDetectionChannel::~ForegroundDetectionChannel() 00064 { } 00065 00066 // ###################################################################### 00067 bool ForegroundDetectionChannel::outputAvailable() const 00068 { return itsMap.initialized(); } 00069 00070 // ###################################################################### 00071 uint ForegroundDetectionChannel::numSubmaps() const 00072 { 00073 return 1; 00074 } 00075 00076 // ###################################################################### 00077 Dims ForegroundDetectionChannel::getMapDims() const 00078 { 00079 if (!this->hasInput()) 00080 LFATAL("Oops! I haven't received any input yet"); 00081 00082 const Dims indims = this->getInputDims(); 00083 00084 return Dims(indims.w() >> itsLevelSpec.getVal().mapLevel(), 00085 indims.h() >> itsLevelSpec.getVal().mapLevel()); 00086 00087 } 00088 00089 // ###################################################################### 00090 void ForegroundDetectionChannel::getFeatures(const Point2D<int>& locn, 00091 std::vector<float>& mean) const 00092 { 00093 LFATAL("not implemented"); 00094 } 00095 00096 // ###################################################################### 00097 void ForegroundDetectionChannel::getFeaturesBatch(std::vector<Point2D<int>*> *locn, 00098 std::vector<std::vector<float> > *mean, 00099 int *count) const 00100 { 00101 LFATAL("not implemented"); 00102 } 00103 00104 00105 // ###################################################################### 00106 void ForegroundDetectionChannel::doInput(const InputFrame& inframe) 00107 { 00108 00109 #ifdef HAVE_OPENCV 00110 ASSERT(inframe.colorByte().initialized()); 00111 LINFO("Input to Foreground Detection Channel ok."); 00112 00113 //Convert the input frame to opencv format 00114 IplImage* inFrame_cv = img2ipl(inframe.colorByte()); 00115 00116 //If the statistics model has not been created (i.e. this is the first frame), 00117 //then create it. 00118 if(itsStatModel_cv == NULL) 00119 itsStatModel_cv = cvCreateFGDStatModel( inFrame_cv ); 00120 00121 //Update the statistics model 00122 cvUpdateBGStatModel( inFrame_cv, itsStatModel_cv ); 00123 00124 //Assign the foreground and background maps 00125 //OpenCV clears out the foreground and background memory at the beginning 00126 //of every icvUpdateFGDStatModel, so let's do a deep copy of the image 00127 //data just to be safe. 00128 //Also, because the source image is byte-valued, let's divide by 255 to get 00129 //a true probability 00130 itsForegroundMap = ipl2gray(itsStatModel_cv->foreground).deepcopy() / 255.0; 00131 00132 //Rescale the image to the correct dimensions 00133 itsMap = rescale(itsForegroundMap, this->getMapDims()); 00134 00135 float mi, ma; 00136 getMinMax(itsMap,mi, ma); 00137 LINFO("FOREGROUND MAP RANGE: [%f .. %f]", mi, ma); 00138 00139 //Free the memory allocated to the input frame - OpenCV makes it's own deep 00140 //copy of this data internally. 00141 cvReleaseImage( &inFrame_cv ); 00142 00143 #endif 00144 00145 00146 } 00147 00148 // ###################################################################### 00149 Image<float> ForegroundDetectionChannel::getSubmap(const uint index) const 00150 { 00151 if (index != 0) 00152 LFATAL("got submap index = %u, but I have only one submap", index); 00153 00154 return itsMap; 00155 } 00156 00157 // ###################################################################### 00158 Image<float> ForegroundDetectionChannel::getRawCSmap(const uint idx) const 00159 { 00160 return Image<float>(); 00161 } 00162 00163 // ###################################################################### 00164 std::string ForegroundDetectionChannel::getSubmapName(const uint index) const 00165 { 00166 return std::string("ForegroundOutput"); 00167 } 00168 00169 00170 // ###################################################################### 00171 Image<float> ForegroundDetectionChannel::getOutput() 00172 { return itsMap; } 00173 00174 00175 // ###################################################################### 00176 /* So things look consistent in everyone's emacs... */ 00177 /* Local Variables: */ 00178 /* indent-tabs-mode: nil */ 00179 /* End: */