Extremal Region Text Detection-ERDetect-Demo

/*M/////////////////////////////////////////////////////////////////////////////////////// // // IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING. // // By downloading, copying, installing or using the software you agree to this license. // If you do not agree to this license, do not download, install, // copy or use the software. // // // License Agreement // For Open Source Computer Vision Library // // Copyright (C) 2000-2008, Intel Corporation, all rights reserved. // Copyright (C) 2009, Willow Garage Inc., all rights reserved. // Third party copyrights are property of their respective owners. // // Redistribution and use in source and binary forms, with or without modification, // are permitted provided that the following conditions are met: // // * Redistribution's of source code must retain the above copyright notice, // this list of conditions and the following disclaimer. // // * Redistribution's in binary form must reproduce the above copyright notice, // this list of conditions and the following disclaimer in the documentation // and/or other materials provided with the distribution. // // * The name of the copyright holders may not be used to endorse or promote products // derived from this software without specific prior written permission. // // This software is provided by the copyright holders and contributors "as is" and // any express or implied warranties, including, but not limited to, the implied // warranties of merchantability and fitness for a particular purpose are disclaimed. // In no event shall the Intel Corporation or contributors be liable for any direct, // indirect, incidental, special, exemplary, or consequential damages // (including, but not limited to, procurement of substitute goods or services; // loss of use, data, or profits; or business interruption) however caused // and on any theory of liability, whether in contract, strict liability, // or tort (including negligence or otherwise) arising in any way out of // the use of this software, even if advised of the possibility of such damage. // //M*/ #include "precomp.hpp" #include "opencv2/imgproc.hpp" #include "opencv2/ml.hpp" #include <limits> #include <fstream> #include <queue> #if defined _MSC_VER && _MSC_VER == 1500 typedef int int_fast32_t; #else #ifndef INT32_MAX #define __STDC_LIMIT_MACROS #include <stdint.h> #endif #endif namespace cv { namespace text { using namespace cv::ml; using namespace std; using namespace cv::ml; // Deletes a tree of ERStat regions starting at root. Used only // internally to this implementation. static void deleteERStatTree(ERStat* root) { queue<ERStat*> to_delete; to_delete.push(root); while (!to_delete.empty()) { ERStat* n = to_delete.front(); to_delete.pop(); ERStat* c = n->child; if (c != NULL) { to_delete.push(c); ERStat* sibling = c->next; while (sibling != NULL) { to_delete.push(sibling); sibling = sibling->next; } } delete n; } } ERStat::ERStat(int init_level, int init_pixel, int init_x, int init_y) : pixel(init_pixel), level(init_level), area(0), perimeter(0), euler(0), probability(1.0), parent(0), child(0), next(0), prev(0), local_maxima(0), max_probability_ancestor(0), min_probability_ancestor(0) { rect = Rect(init_x,init_y,1,1); raw_moments[0] = 0.0; raw_moments[1] = 0.0; central_moments[0] = 0.0; central_moments[1] = 0.0; central_moments[2] = 0.0; crossings = new deque<int>(); crossings->push_back(0); } // derivative classes // the class implementing the interface for the 1st and 2nd stages of Neumann and Matas algorithm class CV_EXPORTS ERFilterNM : public ERFilter { public: //Constructor ERFilterNM(); //Destructor ~ERFilterNM() {} float minProbability; bool nonMaxSuppression; float minProbabilityDiff; // the key method. Takes image on input, vector of ERStat is output for the first stage, // input/output - for the second one. void run( InputArray image, vector<ERStat>& regions ); protected: int thresholdDelta; float maxArea; float minArea; Ptr<ERFilter::Callback> classifier; // count of the rejected/accepted regions int num_rejected_regions; int num_accepted_regions; public: // set/get methods to set the algorithm properties, void setCallback(const Ptr<ERFilter::Callback>& cb); void setThresholdDelta(int thresholdDelta); void setMinArea(float minArea); void setMaxArea(float maxArea); void setMinProbability(float minProbability); void setMinProbabilityDiff(float minProbabilityDiff); void setNonMaxSuppression(bool nonMaxSuppression); int getNumRejected(); private: // pointer to the input/output regions vector vector<ERStat> *regions; // image mask used for feature calculations Mat region_mask; // extract the component tree and store all the ER regions void er_tree_extract( InputArray image ); // accumulate a pixel into an ER void er_add_pixel( ERStat *parent, int x, int y, int non_boundary_neighbours, int non_boundary_neighbours_horiz, int d_C1, int d_C2, int d_C3 ); // merge an ER with its nested parent void er_merge( ERStat *parent, ERStat *child ); // copy extracted regions into the output vector ERStat* er_save( ERStat *er, ERStat *parent, ERStat *prev ); // recursively walk the tree and filter (remove) regions using the callback classifier ERStat* er_tree_filter( InputArray image, ERStat *stat, ERStat *parent, ERStat *prev ); // recursively walk the tree selecting only regions with local maxima probability ERStat* er_tree_nonmax_suppression( ERStat *er, ERStat *parent, ERStat *prev ); }; // default 1st stage classifier class CV_EXPORTS ERClassifierNM1 : public ERFilter::Callback { public: //Constructor ERClassifierNM1(const string& filename); // Destructor ~ERClassifierNM1() {} // The classifier must return probability measure for the region. double eval(const ERStat& stat); private: Ptr<Boost> boost; }; // default 2nd stage classifier class CV_EXPORTS ERClassifierNM2 : public ERFilter::Callback { public: //constructor ERClassifierNM2(const string& filename); // Destructor ~ERClassifierNM2() {} // The classifier must return probability measure for the region. double eval(const ERStat& stat); private: Ptr<Boost> boost; }; // default constructor ERFilterNM::ERFilterNM() { thresholdDelta = 1; minArea = 0.; maxArea = 1.; minProbability = 0.; nonMaxSuppression = false; minProbabilityDiff = 1.; num_accepted_regions = 0; num_rejected_regions = 0; } // the key method. Takes image on input, vector of ERStat is output for the first stage, // input/output for the second one. void ERFilterNM::run( InputArray image, vector<ERStat>& _regions ) { // assert correct image type CV_Assert( image.getMat().type() == CV_8UC1 ); regions = &_regions; region_mask = Mat::zeros(image.getMat().rows+2, image.getMat().cols+2, CV_8UC1); // if regions vector is empty we must extract the entire component tree if ( regions->size() == 0 ) { er_tree_extract( image ); if (nonMaxSuppression) { vector<ERStat> aux_regions; regions->swap(aux_regions); regions->reserve(aux_regions.size()); er_tree_nonmax_suppression( &aux_regions.front(), NULL, NULL ); aux_regions.clear(); } } else // if regions vector is already filled we'll just filter the current regions { // the tree root must have no parent CV_Assert( regions->front().parent == NULL ); vector<E