refactors OCRTesseract interface

modules/text/include/opencv2/text/ocr.hpp

@@ -44,24 +44,15 @@
 #ifndef __OPENCV_TEXT_OCR_HPP__
 #define __OPENCV_TEXT_OCR_HPP__
 
-#include "text_config.hpp"
-
-#ifdef HAVE_TESSERACT
-#include <tesseract/baseapi.h>
-#include <tesseract/resultiterator.h>
-#endif
-
-#include "opencv2/core.hpp"
 #include <vector>
 #include <string>
-
+using namespace std;
 
 namespace cv
 {
 namespace text
 {
 
-using namespace std;
 
 enum
 {
@@ -69,40 +60,35 @@ enum
     OCR_LEVEL_TEXTLINE
 };
 
-#ifdef HAVE_TESSERACT
-class CV_EXPORTS OCRTesseract
+//base class BaseOCR declares a common API that would be used in a typical text recognition scenario
+class CV_EXPORTS BaseOCR
 {
-private:
-    tesseract::TessBaseAPI tess;
-
 public:
-    //Default constructor
-    OCRTesseract(const char* datapath=NULL, const char* language=NULL, const char* char_whitelist=NULL,
-                 tesseract::OcrEngineMode oem=tesseract::OEM_DEFAULT, tesseract::PageSegMode psmode=tesseract::PSM_AUTO);
-
-    ~OCRTesseract();
-
-    void run(Mat& image, string& output_text, vector<Rect>* component_rects=NULL,
-             vector<string>* component_texts=NULL, vector<float>* component_confidences=NULL,
-             int component_level=0);
+    virtual ~BaseOCR() {};
+    virtual void run(Mat& image, string& output_text, vector<Rect>* component_rects=NULL,
+            vector<string>* component_texts=NULL, vector<float>* component_confidences=NULL,
+            int component_level=0) = 0;
 };
-#else
-//stub
-class CV_EXPORTS OCRTesseract
+
+class CV_EXPORTS OCRTesseract : public BaseOCR
 {
 public:
-    //Default constructor
-    OCRTesseract(const char* datapath=NULL, const char* language=NULL, const char* char_whitelist=NULL,
-                 int oem=0, int psmode=0);
-
-    ~OCRTesseract();
-
-    void run(Mat& image, string& output_text, vector<Rect>* component_rects=NULL,
-             vector<string>* component_texts=NULL, vector<float>* component_confidences=NULL,
-             int component_level=0);
+    virtual void run(Mat& image, string& output_text, vector<Rect>* component_rects=NULL,
+            vector<string>* component_texts=NULL, vector<float>* component_confidences=NULL,
+            int component_level=0)
+    {
+      CV_Assert( (image.type() == CV_8UC1) || (image.type() == CV_8UC1) );
+      CV_Assert( (component_level == OCR_LEVEL_TEXTLINE) || (component_level == OCR_LEVEL_WORD) );
+      output_text.clear();
+      if (component_rects != NULL)
+        component_rects->clear();
+      if (component_texts != NULL)
+        component_texts->clear();
+      if (component_confidences != NULL)
+        component_confidences->clear();
+    }
+    static Ptr<OCRTesseract> create(const char* datapath=NULL, const char* language=NULL, const char* char_whitelist=NULL, int oem=3, int psmode=3);
 };
-#endif
-
 
 
 }

modules/text/samples/end_to_end_recognition.cpp

@@ -102,7 +102,7 @@ int main(int argc, char* argv[])
     /*Text Recognition (OCR)*/
 
     double t_r = (double)getTickCount();
-    OCRTesseract* ocr = new OCRTesseract();
+    Ptr<OCRTesseract> ocr = OCRTesseract::create();
     cout << "TIME_OCR_INITIALIZATION = " << ((double)getTickCount() - t_r)*1000/getTickFrequency() << endl;
     string output;
 

modules/text/samples/webcam_demo.cpp

@@ -54,12 +54,12 @@ private:
     vector< vector<Rect> > &boxes;
     vector< vector<string> > &words;
     vector< vector<float> > &confidences;
-    vector< OCRTesseract* > &ocrs;
+    vector< Ptr<OCRTesseract> > &ocrs;
 
 public:
     Parallel_OCR(vector<Mat> &_detections, vector<string> &_outputs, vector< vector<Rect> > &_boxes,
                  vector< vector<string> > &_words, vector< vector<float> > &_confidences, 
-                 vector< OCRTesseract* > &_ocrs)
+                 vector< Ptr<OCRTesseract> > &_ocrs)
         : detections(_detections), outputs(_outputs), boxes(_boxes), words(_words), 
           confidences(_confidences), ocrs(_ocrs)
     {}
@@ -120,11 +120,10 @@ int main(int argc, char* argv[])
 
     //Initialize OCR engine (we initialize 10 instances in order to work several recognitions in parallel)
     int num_ocrs = 10;
-    vector<OCRTesseract*> ocrs;
+    vector< Ptr<OCRTesseract> > ocrs;
     for (int o=0; o<num_ocrs; o++)
     {
-      OCRTesseract* ocr = new OCRTesseract();
-      ocrs.push_back(ocr);
+      ocrs.push_back(OCRTesseract::create());
     }
 
     //cout << "TIME_OCR_INITIALIZATION_ALT = "<< ((double)getTickCount() - t_r)*1000/getTickFrequency() << endl;

modules/text/src/ocr.cpp

@@ -47,25 +47,32 @@
 #include <iostream>
 #include <fstream>
 #include <queue>
+using namespace std;
 
 namespace cv
 {
 namespace text
 {
 
-using namespace std;
-
 
-#ifdef HAVE_TESSERACT
-//Default constructor
-OCRTesseract::OCRTesseract(const char* datapath, const char* language, const char* char_whitelist, tesseract::OcrEngineMode oemode, tesseract::PageSegMode psmode)
+class OCRTesseractImpl : public OCRTesseract
 {
+private:
+#ifdef HAVE_TESSERACT
+    tesseract::TessBaseAPI tess;
+#endif
 
+public:
+    //Default constructor
+    OCRTesseractImpl(const char* datapath, const char* language, const char* char_whitelist, int oemode, int psmode)
+    {
+
+#ifdef HAVE_TESSERACT
     const char *lang = "eng";
     if (language != NULL)
         lang = language;
 
-    if (tess.Init(datapath, lang, oemode))
+    if (tess.Init(datapath, lang, (tesseract::OcrEngineMode)oemode))
     {
         cout << "OCRTesseract: Could not initialize tesseract." << endl;
         throw 1;
@@ -73,7 +80,7 @@ OCRTesseract::OCRTesseract(const char* datapath, const char* language, const cha
 
     //cout << "OCRTesseract: tesseract version " << tess.Version() << endl;
 
-    tesseract::PageSegMode pagesegmode = psmode;
+    tesseract::PageSegMode pagesegmode = (tesseract::PageSegMode)psmode;
     tess.SetPageSegMode(pagesegmode);
 
     if(char_whitelist != NULL)
@@ -82,18 +89,33 @@ OCRTesseract::OCRTesseract(const char* datapath, const char* language, const cha
         tess.SetVariable("tessedit_char_whitelist", "0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ");
 
     tess.SetVariable("save_best_choices", "T");
+#else
+    cout << "OCRTesseract("<<oemode<<psmode<<"): Tesseract not found." << endl;
+    if (datapath != NULL)
+      cout << "            " << datapath << endl;
+    if (language != NULL)
+      cout << "            " << language << endl;
+    if (char_whitelist != NULL)
+      cout << "            " << char_whitelist << endl;
+#endif
+    }
 
-}
+    ~OCRTesseractImpl()
+    {
+#ifdef HAVE_TESSERACT
+      tess.End();
+#endif
+    }
 
-OCRTesseract::~OCRTesseract()
-{
-    tess.End();
-}
+    void run(Mat& image, string& output, vector<Rect>* component_rects=NULL,
+             vector<string>* component_texts=NULL, vector<float>* component_confidences=NULL,
+             int component_level=0)
+    {
+    
+      CV_Assert( (image.type() == CV_8UC1) || (image.type() == CV_8UC1) );
+
+#ifdef HAVE_TESSERACT
 
-void OCRTesseract::run(Mat& image, string& output, vector<Rect>* component_rects,
-                       vector<string>* component_texts, vector<float>* component_confidences, int component_level)
-{
-    CV_Assert( (image.type() == CV_8UC1) || (image.type() == CV_8UC1) );
     if (component_texts != 0)
         component_texts->clear();
     if (component_rects != 0)
@@ -135,30 +157,8 @@ void OCRTesseract::run(Mat& image, string& output, vector<Rect>* component_rects
     }
 
     tess.Clear();
-}
-#else
-//Stub constructor
-OCRTesseract::OCRTesseract(const char* datapath, const char* language, const char* char_whitelist, int oemode, int psmode)
-{
-    cout << "OCRTesseract("<<oemode<<psmode<<"): Tesseract not found." << endl;
-    if (datapath != NULL)
-      cout << "            " << datapath << endl;
-    if (language != NULL)
-      cout << "            " << language << endl;
-    if (char_whitelist != NULL)
-      cout << "            " << char_whitelist << endl;
-}
 
-//Stub destructor
-OCRTesseract::~OCRTesseract()
-{
-}
-
-//Stub method, does nothing
-void OCRTesseract::run(Mat& image, string& output, vector<Rect>* component_rects,
-                       vector<string>* component_texts, vector<float>* component_confidences, int component_level)
-{
-    CV_Assert( (image.type() == CV_8UC1) || (image.type() == CV_8UC1) );
+#else
 
     cout << "OCRTesseract(" << component_level << image.type() <<"): Tesseract not found." << endl;
     output.clear();
@@ -168,9 +168,16 @@ void OCRTesseract::run(Mat& image, string& output, vector<Rect>* component_rects
         component_texts->clear();
     if(component_confidences)
         component_confidences->clear();
-}
 #endif
+    }
+
 
+};
+
+Ptr<OCRTesseract> OCRTesseract::create(const char* datapath, const char* language, const char* char_whitelist, int oem, int psmode)
+{
+   return makePtr<OCRTesseractImpl>(datapath,language,char_whitelist,oem,psmode);
+}
 
 
 }

modules/text/src/precomp.hpp

@@ -45,4 +45,11 @@
 
 #include "opencv2/text.hpp"
 
+#include "text_config.hpp"
+
+#ifdef HAVE_TESSERACT
+#include <tesseract/baseapi.h>
+#include <tesseract/resultiterator.h>
+#endif
+
 #endif