Point Cloud Library (PCL)
1.7.0
|
00001 /* 00002 * Software License Agreement (BSD License) 00003 * 00004 * Point Cloud Library (PCL) - www.pointclouds.org 00005 * Copyright (c) 2013-, Open Perception, Inc. 00006 * 00007 * All rights reserved. 00008 * 00009 * Redistribution and use in source and binary forms, with or without 00010 * modification, are permitted provided that the following conditions 00011 * are met: 00012 * 00013 * * Redistributions of source code must retain the above copyright 00014 * notice, this list of conditions and the following disclaimer. 00015 * * Redistributions in binary form must reproduce the above 00016 * copyright notice, this list of conditions and the following 00017 * disclaimer in the documentation and/or other materials provided 00018 * with the distribution. 00019 * * Neither the name of the copyright holder(s) nor the names of its 00020 * contributors may be used to endorse or promote products derived 00021 * from this software without specific prior written permission. 00022 * 00023 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 00024 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT 00025 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS 00026 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE 00027 * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, 00028 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, 00029 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; 00030 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER 00031 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 00032 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN 00033 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 00034 * POSSIBILITY OF SUCH DAMAGE. 00035 * 00036 * person_classifier.h 00037 * Created on: Nov 30, 2012 00038 * Author: Matteo Munaro 00039 */ 00040 00041 #ifndef PCL_PEOPLE_PERSON_CLASSIFIER_H_ 00042 #define PCL_PEOPLE_PERSON_CLASSIFIER_H_ 00043 00044 #include <pcl/people/person_cluster.h> 00045 #include <pcl/people/hog.h> 00046 00047 namespace pcl 00048 { 00049 namespace people 00050 { 00051 template <typename PointT> class PersonClassifier; 00052 00053 template <typename PointT> 00054 class PersonClassifier 00055 { 00056 protected: 00057 00058 /** \brief Height of the image patch to classify. */ 00059 int window_height_; 00060 00061 /** \brief Width of the image patch to classify. */ 00062 int window_width_; 00063 00064 /** \brief SVM offset. */ 00065 float SVM_offset_; 00066 00067 /** \brief SVM weights vector. */ 00068 std::vector<float> SVM_weights_; 00069 00070 public: 00071 00072 typedef pcl::PointCloud<PointT> PointCloud; 00073 typedef boost::shared_ptr<PointCloud> PointCloudPtr; 00074 00075 /** \brief Constructor. */ 00076 PersonClassifier (); 00077 00078 /** \brief Destructor. */ 00079 virtual ~PersonClassifier (); 00080 00081 /** \brief Load SVM parameters from a text file. 00082 * 00083 * \param[in] svm_filename Filename containing SVM parameters. 00084 * 00085 * \return true if SVM has been correctly set, false otherwise. 00086 */ 00087 bool 00088 loadSVMFromFile (std::string svm_filename); 00089 00090 /** 00091 * \brief Set trained SVM for person confidence estimation. 00092 * 00093 * \param[in] window_height Detection window height. 00094 * \param[in] window_width Detection window width. 00095 * \param[in] SVM_weights SVM weights vector. 00096 * \param[in] SVM_offset SVM offset. 00097 */ 00098 void 00099 setSVM (int window_height, int window_width, std::vector<float> SVM_weights, float SVM_offset); 00100 00101 /** 00102 * \brief Get trained SVM for person confidence estimation. 00103 * 00104 * \param[out] window_height Detection window height. 00105 * \param[out] window_width Detection window width. 00106 * \param[out] SVM_weights SVM weights vector. 00107 * \param[out] SVM_offset SVM offset. 00108 */ 00109 void 00110 getSVM (int& window_height, int& window_width, std::vector<float>& SVM_weights, float& SVM_offset); 00111 00112 /** 00113 * \brief Resize an image represented by a pointcloud containing RGB information. 00114 * 00115 * \param[in] input_image A pointer to a pointcloud containing RGB information. 00116 * \param[out] output_image A pointer to the output pointcloud. 00117 * \param[in] width Output width. 00118 * \param[in] height Output height. 00119 */ 00120 void 00121 resize (PointCloudPtr& input_image, PointCloudPtr& output_image, 00122 int width, int height); 00123 00124 /** 00125 * \brief Copies an image and makes a black border around it, where the source image is not present. 00126 * 00127 * \param[in] input_image A pointer to a pointcloud containing RGB information. 00128 * \param[out] output_image A pointer to the output pointcloud. 00129 * \param[in] xmin x coordinate of the top-left point of the bbox to copy from the input image. 00130 * \param[in] ymin y coordinate of the top-left point of the bbox to copy from the input image. 00131 * \param[in] width Output width. 00132 * \param[in] height Output height. 00133 */ 00134 void 00135 copyMakeBorder (PointCloudPtr& input_image, PointCloudPtr& output_image, 00136 int xmin, int ymin, int width, int height); 00137 00138 /** 00139 * \brief Classify the given portion of image. 00140 * 00141 * \param[in] height The height of the image patch to classify, in pixels. 00142 * \param[in] xc The x-coordinate of the center of the image patch to classify, in pixels. 00143 * \param[in] yc The y-coordinate of the center of the image patch to classify, in pixels. 00144 * \param[in] image The whole image (pointer to a point cloud containing RGB information) containing the object to classify. 00145 * \return The classification score given by the SVM. 00146 */ 00147 double 00148 evaluate (float height, float xc, float yc, PointCloudPtr& image); 00149 00150 /** 00151 * \brief Compute person confidence for a given PersonCluster. 00152 * 00153 * \param[in] image The input image (pointer to a point cloud containing RGB information). 00154 * \param[in] bottom Theoretical bottom point of the cluster projected to the image. 00155 * \param[in] top Theoretical top point of the cluster projected to the image. 00156 * \param[in] centroid Theoretical centroid point of the cluster projected to the image. 00157 * \param[in] vertical If true, the sensor is considered to be vertically placed (portrait mode). 00158 * \return The person confidence. 00159 */ 00160 double 00161 evaluate (PointCloudPtr& image, Eigen::Vector3f& bottom, Eigen::Vector3f& top, Eigen::Vector3f& centroid, 00162 bool vertical); 00163 }; 00164 } /* namespace people */ 00165 } /* namespace pcl */ 00166 #include <pcl/people/impl/person_classifier.hpp> 00167 #endif /* PCL_PEOPLE_PERSON_CLASSIFIER_H_ */