00001 /*!@file HMAX/test-hmaxRec.C aginst object recognition */ 00002 00003 // //////////////////////////////////////////////////////////////////// // 00004 // The iLab Neuromorphic Vision C++ Toolkit - Copyright (C) 2001 by the // 00005 // University of Southern California (USC) and the iLab at USC. // 00006 // See http://iLab.usc.edu for information about this project. // 00007 // //////////////////////////////////////////////////////////////////// // 00008 // Major portions of the iLab Neuromorphic Vision Toolkit are protected // 00009 // under the U.S. patent ``Computation of Intrinsic Perceptual Saliency // 00010 // in Visual Environments, and Applications'' by Christof Koch and // 00011 // Laurent Itti, California Institute of Technology, 2001 (patent // 00012 // pending; application number 09/912,225 filed July 23, 2001; see // 00013 // http://pair.uspto.gov/cgi-bin/final/home.pl for current status). // 00014 // //////////////////////////////////////////////////////////////////// // 00015 // This file is part of the iLab Neuromorphic Vision C++ Toolkit. // 00016 // // 00017 // The iLab Neuromorphic Vision C++ Toolkit is free software; you can // 00018 // redistribute it and/or modify it under the terms of the GNU General // 00019 // Public License as published by the Free Software Foundation; either // 00020 // version 2 of the License, or (at your option) any later version. // 00021 // // 00022 // The iLab Neuromorphic Vision C++ Toolkit is distributed in the hope // 00023 // that it will be useful, but WITHOUT ANY WARRANTY; without even the // 00024 // implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR // 00025 // PURPOSE. See the GNU General Public License for more details. // 00026 // // 00027 // You should have received a copy of the GNU General Public License // 00028 // along with the iLab Neuromorphic Vision C++ Toolkit; if not, write // 00029 // to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, // 00030 // Boston, MA 02111-1307 USA. // 00031 // //////////////////////////////////////////////////////////////////// // 00032 // 00033 // Primary maintainer for this file: Lior Elazary 00034 // $HeadURL: svn://isvn.usc.edu/software/invt/trunk/saliency/src/HMAX/test-hmaxRec.C $ 00035 // $Id: test-hmaxRec.C 9412 2008-03-10 23:10:15Z farhan $ 00036 // 00037 00038 #include "HMAX/Hmax.H" 00039 #include "Image/Image.H" 00040 #include "Image/ColorOps.H" 00041 #include "Image/MathOps.H" 00042 #include "Image/CutPaste.H" 00043 #include "Image/Transforms.H" 00044 #include "Raster/Raster.H" 00045 #include "Util/Types.H" 00046 #include "Util/log.H" 00047 #include "GUI/DebugWin.H" 00048 #include "Media/TestImages.H" 00049 00050 #include <iostream> 00051 #include <unistd.h> 00052 00053 // number of orientations to use in Hmax 00054 #define NORI 4 00055 00056 int main(const int argc, const char **argv) 00057 { 00058 int debug=0; 00059 if (argc != 2) 00060 { std::cerr<<"USAGE: test-hmaxRec <file>"<<std::endl; exit(1); } 00061 00062 // get an Hmax object: 00063 std::vector<int> scss(5); 00064 scss[0] = 0; scss[1] = 2; scss[2] = 5; scss[3] = 8; scss[4] = 12; 00065 std::vector<int> spss(4); 00066 spss[0] = 4; spss[1] = 6; spss[2] = 9; spss[3] = 12; 00067 Hmax hmax(NORI, spss, scss); 00068 00069 const char *imageSetFile = argv[1]; 00070 00071 //load the images 00072 TestImages testImages(imageSetFile, TestImages::XMLFILE); 00073 00074 //int numMatches = 0; //the number of correct matches 00075 int totalObjects = 0; //the number of objects presented to the network 00076 00077 for (uint scene=0; scene<testImages.getNumScenes(); scene++) //look at all the scenes 00078 { 00079 00080 for (uint obj=0; obj<testImages.getNumObj(scene); obj++) //look at all the objects 00081 { 00082 TestImages::ObjData objData = testImages.getObjectData(scene, obj); 00083 00084 //crop the obj image to 256x256; 00085 Point2D<int> upperLeft(0,0); 00086 00087 if (objData.img.getWidth() > 256 || objData.img.getHeight() > 256) 00088 upperLeft = Point2D<int>((objData.img.getWidth()/2)-128, (objData.img.getHeight()/2)-128); 00089 00090 LINFO("Upper left %ix%i", upperLeft.i, upperLeft.j); 00091 Image<PixRGB<byte> > inputImg = crop(objData.img, upperLeft, Dims(256,256), true); 00092 //center the object 00093 if (objData.img.getWidth() < 256 || objData.img.getHeight() < 256) 00094 inputImg = shift(inputImg, 128-(objData.img.getWidth()/2), 128-(objData.img.getHeight()/2)); 00095 00096 totalObjects++; 00097 00098 // read the image: 00099 Image<byte> input = luminance(inputImg); //convert to gray 00100 00101 Image<PixRGB<byte> > out = toRGB(input); 00102 if(debug) SHOWIMG(out); 00103 00104 00105 // pass image through Hmax model: 00106 Image<float> inputf = input; // convert image to floats 00107 Image<float> c2resp = hmax.getC2(inputf); 00108 00109 float mi, ma; getMinMax(c2resp, mi, ma); 00110 LINFO("%ix%i min=%f max=%f", c2resp.getWidth(), c2resp.getHeight(), mi, ma); 00111 00112 printf("%i ", objData.id); 00113 for(int i=0; i<c2resp.getWidth()*c2resp.getHeight(); i++) 00114 { 00115 printf("%i:%f ", i,c2resp[i]); 00116 } 00117 printf("\n"); 00118 00119 00120 //use svm to classify 00121 00122 // display C2 response in an X window: 00123 if (debug) 00124 { 00125 c2resp = scaleBlock(c2resp, input.getDims()); 00126 00127 inplaceNormalize(c2resp, 0.0F, 255.0F); 00128 SHOWIMG(c2resp); 00129 } 00130 } 00131 } 00132 00133 return 0; 00134 } 00135 00136 // ###################################################################### 00137 /* So things look consistent in everyone's emacs... */ 00138 /* Local Variables: */ 00139 /* indent-tabs-mode: nil */ 00140 /* End: */