TessBaseAPI Class Reference

#include <baseapi.h>

List of all members.

---

Detailed Description

Definition at line 35 of file baseapi.h.

Static Public Member Functions

• static void ClearAdaptiveClassifier ()
• static void DumpPGM (const char *filename)
• static void End ()
• static int Init (const char *datapath, const char *outputbase, const char *configfile, bool numeric_mode, int argc, char *argv[])
• static char * TesseractRect (const UINT8 *imagedata, int bytes_per_pixel, int bytes_per_line, int left, int top, int width, int height)

Static Protected Member Functions

• static void CopyBinaryRect (const UINT8 *imagedata, int bytes_per_line, int left, int top, int width, int height)
• static void CopyImageToTesseract (const UINT8 *imagedata, int bytes_per_pixel, int bytes_per_line, int left, int top, int width, int height)
• static void FindLines (BLOCK_LIST *block_list)
• static void HistogramRect (const UINT8 *imagedata, int bytes_per_pixel, int bytes_per_line, int left, int top, int right, int bottom, int *histogram)
• static int OtsuStats (const int *histogram, int *H_out, int *omega0_out)
• static void OtsuThreshold (const UINT8 *imagedata, int bytes_per_pixel, int bytes_per_line, int left, int top, int right, int bottom, int *thresholds, int *hi_values)
• static PAGE_RES * Recognize (BLOCK_LIST *block_list, ETEXT_DESC *monitor)
• static char * RecognizeToString ()
• static char * TesseractToText (PAGE_RES *page_res)
• static void ThresholdRect (const UINT8 *imagedata, int bytes_per_pixel, int bytes_per_line, int left, int top, int width, int height, const int *thresholds, const int *hi_values)

---

Member Function Documentation

void TessBaseAPI::ClearAdaptiveClassifier

(

)

[static]

Definition at line 90 of file baseapi.cpp.

References ResetAdaptiveClassifier().

                                          {
  ResetAdaptiveClassifier();
}

void TessBaseAPI::CopyBinaryRect	(	const UINT8 *	imagedata,
		int	bytes_per_line,
		int	left,
		int	top,
		int	width,
		int	height
	)			[static, protected]

Definition at line 308 of file baseapi.cpp.

References IMAGE::capture(), copy_sub_image(), IMAGE::create(), and page_image.

Referenced by CopyImageToTesseract().

                                                        {
  // Copy binary image, cutting out the required rectangle.
  IMAGE image;
  image.capture(const_cast<UINT8*>(imagedata),
                bytes_per_line*8, top + height, 1);
  page_image.create(width, height, 1);
  copy_sub_image(&image, left, top, width, height, &page_image, 0, 0, false);
}

void TessBaseAPI::CopyImageToTesseract	(	const UINT8 *	imagedata,
		int	bytes_per_pixel,
		int	bytes_per_line,
		int	left,
		int	top,
		int	width,
		int	height
	)			[static, protected]

Definition at line 119 of file baseapi.cpp.

References CopyBinaryRect(), OtsuThreshold(), and ThresholdRect().

Referenced by TesseractRect().

                                                              {
  if (bytes_per_pixel > 0) {
    // Threshold grey or color.
    int* thresholds = new int[bytes_per_pixel];
    int* hi_values = new int[bytes_per_pixel];

    // Compute the thresholds.
    OtsuThreshold(imagedata, bytes_per_pixel, bytes_per_line,
                  left, top, left + width, top + height,
                  thresholds, hi_values);

    // Threshold the image to the tesseract global page_image.
    ThresholdRect(imagedata, bytes_per_pixel, bytes_per_line,
                  left, top, width, height,
                  thresholds, hi_values);
    delete [] thresholds;
    delete [] hi_values;
  } else {
    CopyBinaryRect(imagedata, bytes_per_line, left, top, width, height);
  }
}

void TessBaseAPI::DumpPGM

(

const char *

filename

)

[static]

Definition at line 101 of file baseapi.cpp.

References IMAGE::get_line(), IMAGE::get_xsize(), IMAGE::get_ysize(), IMAGELINE::init(), INT32FORMAT, page_image, and IMAGELINE::pixels.

                                              {
  IMAGELINE line;
  line.init(page_image.get_xsize());
  FILE *fp = fopen(filename, "w");
  fprintf(fp, "P5 " INT32FORMAT " " INT32FORMAT " 255\n", page_image.get_xsize(),
          page_image.get_ysize());
  for (int j = page_image.get_ysize()-1; j >= 0 ; --j) {
    page_image.get_line(0, j, page_image.get_xsize(), &line, 0);
    for (int i = 0; i < page_image.get_xsize(); ++i) {
      UINT8 b = line.pixels[i] ? 255 : 0;
      fwrite(&b, 1, 1, fp);
    }
  }
  fclose(fp);
}

void TessBaseAPI::End

(

)

[static]

Definition at line 95 of file baseapi.cpp.

References end_tesseract(), and ResetAdaptiveClassifier().

Referenced by main().

                      {
  ResetAdaptiveClassifier();
  end_tesseract();
}

void TessBaseAPI::FindLines

(

BLOCK_LIST *

block_list

)

[static, protected]

Definition at line 333 of file baseapi.cpp.

References pgeditor_read_file().

Referenced by RecognizeToString().

                                                  {
  STRING input_file = "noname.tif";
  // The following call creates a full-page block and then runs connected
  // component analysis and text line creation.
  pgeditor_read_file(input_file, block_list);
}

void TessBaseAPI::HistogramRect	(	const UINT8 *	imagedata,
		int	bytes_per_pixel,
		int	bytes_per_line,
		int	left,
		int	top,
		int	right,
		int	bottom,
		int *	histogram
	)			[static, protected]

Definition at line 211 of file baseapi.cpp.

Referenced by OtsuThreshold().

                                                {
  int width = right - left;
  memset(histogram, 0, sizeof(*histogram) * 256);
  const UINT8* pix = imagedata +
                     top*bytes_per_line +
                     left*bytes_per_pixel;
  for (int y = top; y < bottom; ++y) {
    for (int x = 0; x < width; ++x) {
      ++histogram[pix[x * bytes_per_pixel]];
    }
    pix += bytes_per_line;
  }
}

int TessBaseAPI::Init	(	const char *	datapath,
		const char *	outputbase,
		const char *	configfile,
		bool	numeric_mode,
		int	argc,
		char *	argv[]
	)			[static]

Definition at line 52 of file baseapi.cpp.

References init_tesseract().

Referenced by main().

                                              {
  int result = init_tesseract(datapath, outputbase, configfile, argc, argv);
  bln_numericmode.set_value(numeric_mode);
  return result;
}

int TessBaseAPI::OtsuStats	(	const int *	histogram,
		int *	H_out,
		int *	omega0_out
	)			[static, protected]

Definition at line 232 of file baseapi.cpp.

References NULL.

Referenced by OtsuThreshold().

                                            {
  int H = 0;
  double mu_T = 0.0;
  for (int i = 0; i < 256; ++i) {
    H += histogram[i];
    mu_T += i * histogram[i];
  }

  // Now maximize sig_sq_B over t.
  // http://www.ctie.monash.edu.au/hargreave/Cornall_Terry_328.pdf
  int best_t = -1;
  int omega_0, omega_1;
  int best_omega_0 = 0;
  double best_sig_sq_B = 0.0;
  double mu_0, mu_1, mu_t;
  omega_0 = 0;
  mu_t = 0.0;
  for (int t = 0; t < 255; ++t) {
    omega_0 += histogram[t];
    mu_t += t * static_cast<double>(histogram[t]);
    if (omega_0 == 0)
      continue;
    omega_1 = H - omega_0;
    mu_0 = mu_t / omega_0;
    mu_1 = (mu_T - mu_t) / omega_1;
    double sig_sq_B = mu_1 - mu_0;
    sig_sq_B *= sig_sq_B * omega_0 * omega_1;
    if (best_t < 0 || sig_sq_B > best_sig_sq_B) {
      best_sig_sq_B = sig_sq_B;
      best_t = t;
      best_omega_0 = omega_0;
    }
  }
  if (H_out != NULL) *H_out = H;
  if (omega0_out != NULL) *omega0_out = best_omega_0;
  return best_t;
}

void TessBaseAPI::OtsuThreshold	(	const UINT8 *	imagedata,
		int	bytes_per_pixel,
		int	bytes_per_line,
		int	left,
		int	top,
		int	right,
		int	bottom,
		int *	thresholds,
		int *	hi_values
	)			[static, protected]

Definition at line 152 of file baseapi.cpp.

References HistogramRect(), and OtsuStats().

Referenced by CopyImageToTesseract().

                                                {
  // Of all channels with no good hi_value, keep the best so we can always
  // produce at least one answer.
  int best_hi_value = 0;
  int best_hi_index = 0;
  bool any_good_hivalue = false;
  double best_hi_dist = 0.0;

  for (int ch = 0; ch < bytes_per_pixel; ++ch) {
    thresholds[ch] = 0;
    hi_values[ch] = -1;
    // Compute the histogram of the image rectangle.
    int histogram[256];
    HistogramRect(imagedata + ch, bytes_per_pixel, bytes_per_line,
                  left, top, right, bottom, histogram);
    int H;
    int best_omega_0;
    int best_t = OtsuStats(histogram, &H, &best_omega_0);
    // To be a convincing foreground we must have a small fraction of H
    // or to be a convincing background we must have a large fraction of H.
    // In between we assume this channel contains no thresholding information.
    int hi_value = best_omega_0 < H * 0.5;
    thresholds[ch] = best_t;
    if (best_omega_0 > H * 0.75) {
      any_good_hivalue = true;
      hi_values[ch] = 0;
    }
    else if (best_omega_0 < H * 0.25) {
      any_good_hivalue = true;
      hi_values[ch] = 1;
    }
    else {
      // In case all channels are like this, keep the best of the bad lot.
      double hi_dist = hi_value ? (H - best_omega_0) : best_omega_0;
      if (hi_dist > best_hi_dist) {
        best_hi_dist = hi_dist;
        best_hi_value = hi_value;
        best_hi_index = ch;
      }
    }
  }
  if (!any_good_hivalue) {
    // Use the best of the ones that were not good enough.
    hi_values[best_hi_index] = best_hi_value;
  }
}

PAGE_RES * TessBaseAPI::Recognize	(	BLOCK_LIST *	block_list,
		ETEXT_DESC *	monitor
	)			[static, protected]

Definition at line 342 of file baseapi.cpp.

References apply_box_training(), apply_boxes(), pgeditor_main(), recog_all_words(), and start_variables_editor().

Referenced by RecognizeToString().

                                                                            {
  if (tessedit_resegment_from_boxes)
    apply_boxes(block_list);
  if (edit_variables)
    start_variables_editor();

  PAGE_RES* page_res = new PAGE_RES(block_list);
  if (interactive_mode) {
    pgeditor_main(block_list);                  //pgeditor user I/F
  } else if (tessedit_train_from_boxes) {
    apply_box_training(block_list);
  } else {
    // Now run the main recognition.
    recog_all_words(page_res, monitor);
  }
  return page_res;
}

char * TessBaseAPI::RecognizeToString

(

)

[static, protected]

Definition at line 321 of file baseapi.cpp.

References FindLines(), NULL, Recognize(), and TesseractToText().

Referenced by TesseractRect().

                                     {
  BLOCK_LIST    block_list;

  FindLines(&block_list);

  // Now run the main recognition.
  PAGE_RES* page_res = Recognize(&block_list, NULL);

  return TesseractToText(page_res);
}

char * TessBaseAPI::TesseractRect	(	const UINT8 *	imagedata,
		int	bytes_per_pixel,
		int	bytes_per_line,
		int	left,
		int	top,
		int	width,
		int	height
	)			[static]

Definition at line 71 of file baseapi.cpp.

References CopyImageToTesseract(), kMinRectSize, NULL, and RecognizeToString().

Referenced by main().

                                                        {
#ifndef TEXT_VERBOSE
  if (width < kMinRectSize || height < kMinRectSize)
    return NULL;  // Nothing worth doing.
#endif // TEXT_VERBOSE - useful for debugging single characters!

  // Copy/Threshold the image to the tesseract global page_image.
  CopyImageToTesseract(imagedata, bytes_per_pixel, bytes_per_line,
                       left, top, width, height);

  return RecognizeToString();
}

char * TessBaseAPI::TesseractToText

(

PAGE_RES *

page_res

)

[static, protected]

Definition at line 362 of file baseapi.cpp.

References WERD_RES::best_choice, PAGE_RES_IT::forward(), NULL, PAGE_RES_IT::restart_page(), W_EOL, and PAGE_RES_IT::word().

Referenced by RecognizeToString().

                                                     {
  if (page_res != NULL) {
    int total_length = 2;
    PAGE_RES_IT   page_res_it(page_res);
    // Iterate over the data structures to extract the recognition result.
    for (page_res_it.restart_page(); page_res_it.word () != NULL;
         page_res_it.forward()) {
      WERD_RES *word = page_res_it.word();
      WERD_CHOICE* choice = word->best_choice;
      if (choice != NULL) {
        total_length += choice->string().length() + 1;
      }
    }
    char* result = new char[total_length];
    char* ptr = result;
    for (page_res_it.restart_page(); page_res_it.word () != NULL;
         page_res_it.forward()) {
      WERD_RES *word = page_res_it.word();
      WERD_CHOICE* choice = word->best_choice;
      if (choice != NULL) {
        strcpy(ptr, choice->string().string());
        ptr += strlen(ptr);
        if (word->word->flag(W_EOL))
          *ptr++ = '\n';
        else
          *ptr++ = ' ';
      }
    }
    *ptr++ = '\n';
    *ptr = '\0';
    delete page_res;
    return result;
  }
  return NULL;
}

void TessBaseAPI::ThresholdRect	(	const UINT8 *	imagedata,
		int	bytes_per_pixel,
		int	bytes_per_line,
		int	left,
		int	top,
		int	width,
		int	height,
		const int *	thresholds,
		const int *	hi_values
	)			[static, protected]

Definition at line 4 of file api.cpp.

References IMAGE::create(), IMAGELINE::init(), page_image, IMAGELINE::pixels, and IMAGE::put_line().

Referenced by CopyImageToTesseract().

                                                      {
  IMAGELINE line;
  page_image.create(width, height, 1);
  line.init(width);
  // For each line in the image, fill the IMAGELINE class and put it into the
  // Tesseract global page_image. Note that Tesseract stores images with the
  // bottom at y=0 and 0 is black, so we need 2 kinds of inversion.
  //fmg: unless told otherwise, I'll assume that UNIT8* is == char*


  const UINT8* data = imagedata + top*bytes_per_line + left*bytes_per_pixel;
  for (int y = height - 1 ; y >= 0; --y) {
    const UINT8* pix = data;
    for (int x = 0; x < width; ++x, pix += bytes_per_pixel) {
      line.pixels[x] = 1;
      for (int ch = 0; ch < bytes_per_pixel; ++ch) {
        if (hi_values[ch] >= 0 && (pix[ch] > thresholds[ch]) == (hi_values[ch] == 0)) {
          line.pixels[x] = 0;
          break;
        }//if
      }//for bytes in each pixel
    }//for x
    page_image.put_line(0, y, width, &line, 0);
    data += bytes_per_line;
  }//for y
}//TessBaseAPI::ThresholdRect

---

The documentation for this class was generated from the following files:

• ccmain/baseapi.h
• ccmain/api.cpp
• ccmain/baseapi.cpp

---