blobbox.h File Reference

#include <cinttypes>
#include <cmath>
#include <cstdint>
#include "elst.h"
#include "elst2.h"
#include "errcode.h"
#include "ocrblock.h"
#include "params.h"
#include "pdblock.h"
#include "points.h"
#include "quspline.h"
#include "rect.h"
#include "scrollview.h"
#include "statistc.h"
#include "stepblob.h"
#include "tprintf.h"
#include "werd.h"

Go to the source code of this file.

Classes
class	BLOBNBOX
class	TO_ROW
class	TO_BLOCK

Namespaces
	tesseract

Enumerations
enum	PITCH_TYPE {   PITCH_DUNNO, PITCH_DEF_FIXED, PITCH_MAYBE_FIXED, PITCH_DEF_PROP,   PITCH_MAYBE_PROP, PITCH_CORR_FIXED, PITCH_CORR_PROP }
enum	TabType {   TT_NONE, TT_DELETED, TT_MAYBE_RAGGED, TT_MAYBE_ALIGNED,   TT_CONFIRMED, TT_VLINE }
enum	BlobRegionType {   BRT_NOISE, BRT_HLINE, BRT_VLINE, BRT_RECTIMAGE,   BRT_POLYIMAGE, BRT_UNKNOWN, BRT_VERT_TEXT, BRT_TEXT,   BRT_COUNT }
enum	BlobNeighbourDir {   BND_LEFT, BND_BELOW, BND_RIGHT, BND_ABOVE,   BND_COUNT }
enum	BlobSpecialTextType {   BSTT_NONE, BSTT_ITALIC, BSTT_DIGIT, BSTT_MATH,   BSTT_UNCLEAR, BSTT_SKIP, BSTT_COUNT }
enum	BlobTextFlowType {   BTFT_NONE, BTFT_NONTEXT, BTFT_NEIGHBOURS, BTFT_CHAIN,   BTFT_STRONG_CHAIN, BTFT_TEXT_ON_IMAGE, BTFT_LEADER, BTFT_COUNT }

Functions
BlobNeighbourDir	DirOtherWay (BlobNeighbourDir dir)
bool	DominatesInMerge (BlobTextFlowType type1, BlobTextFlowType type2)
void	find_cblob_limits (C_BLOB *blob, float leftx, float rightx, FCOORD rotation, float &ymin, float &ymax)
void	find_cblob_vlimits (C_BLOB *blob, float leftx, float rightx, float &ymin, float &ymax)
void	find_cblob_hlimits (C_BLOB *blob, float bottomy, float topy, float &xmin, float &xymax)
C_BLOB *	crotate_cblob (C_BLOB *blob, FCOORD rotation)
TBOX	box_next (BLOBNBOX_IT *it)
TBOX	box_next_pre_chopped (BLOBNBOX_IT *it)
void	vertical_cblob_projection (C_BLOB *blob, STATS *stats)
void	vertical_coutline_projection (C_OUTLINE *outline, STATS *stats)
void	plot_blob_list (ScrollView *win, BLOBNBOX_LIST *list, ScrollView::Color body_colour, ScrollView::Color child_colour)

Variables
double	textord_error_weight = 3

Enumeration Type Documentation

BlobNeighbourDir

enum BlobNeighbourDir

Enumerator
BND_LEFT
BND_BELOW
BND_RIGHT
BND_ABOVE
BND_COUNT

Definition at line 88 of file blobbox.h.

                      {
  BND_LEFT,
  BND_BELOW,
  BND_RIGHT,
  BND_ABOVE,
  BND_COUNT
};

BlobRegionType

enum BlobRegionType

Enumerator
BRT_NOISE
BRT_HLINE
BRT_VLINE
BRT_RECTIMAGE
BRT_POLYIMAGE
BRT_UNKNOWN
BRT_VERT_TEXT
BRT_TEXT
BRT_COUNT

Definition at line 73 of file blobbox.h.

                    {
  BRT_NOISE,      // Neither text nor image.
  BRT_HLINE,      // Horizontal separator line.
  BRT_VLINE,      // Vertical separator line.
  BRT_RECTIMAGE,  // Rectangular image.
  BRT_POLYIMAGE,  // Non-rectangular image.
  BRT_UNKNOWN,    // Not determined yet.
  BRT_VERT_TEXT,  // Vertical alignment, not necessarily vertically oriented.
  BRT_TEXT,       // Convincing text.

  BRT_COUNT       // Number of possibilities.
};

BlobSpecialTextType

enum BlobSpecialTextType

Enumerator
BSTT_NONE
BSTT_ITALIC
BSTT_DIGIT
BSTT_MATH
BSTT_UNCLEAR
BSTT_SKIP
BSTT_COUNT

Definition at line 97 of file blobbox.h.

                         {
  BSTT_NONE,  // No special.
  BSTT_ITALIC,  // Italic style.
  BSTT_DIGIT,  // Digit symbols.
  BSTT_MATH,  // Mathmatical symobls (not including digit).
  BSTT_UNCLEAR,  // Characters with low recognition rate.
  BSTT_SKIP,  // Characters that we skip labeling (usually too small).
  BSTT_COUNT
};

BlobTextFlowType

enum BlobTextFlowType

Enumerator
BTFT_NONE
BTFT_NONTEXT
BTFT_NEIGHBOURS
BTFT_CHAIN
BTFT_STRONG_CHAIN
BTFT_TEXT_ON_IMAGE
BTFT_LEADER
BTFT_COUNT

Definition at line 115 of file blobbox.h.

                      {
  BTFT_NONE,           // No text flow set yet.
  BTFT_NONTEXT,        // Flow too poor to be likely text.
  BTFT_NEIGHBOURS,     // Neighbours support flow in this direction.
  BTFT_CHAIN,          // There is a weak chain of text in this direction.
  BTFT_STRONG_CHAIN,   // There is a strong chain of text in this direction.
  BTFT_TEXT_ON_IMAGE,  // There is a strong chain of text on an image.
  BTFT_LEADER,         // Leader dots/dashes etc.
  BTFT_COUNT
};

PITCH_TYPE

enum PITCH_TYPE

Enumerator
PITCH_DUNNO
PITCH_DEF_FIXED
PITCH_MAYBE_FIXED
PITCH_DEF_PROP
PITCH_MAYBE_PROP
PITCH_CORR_FIXED
PITCH_CORR_PROP

Definition at line 45 of file blobbox.h.

{
  PITCH_DUNNO,                   // insufficient data
  PITCH_DEF_FIXED,               // definitely fixed
  PITCH_MAYBE_FIXED,             // could be
  PITCH_DEF_PROP,
  PITCH_MAYBE_PROP,
  PITCH_CORR_FIXED,
  PITCH_CORR_PROP
};

TabType

enum TabType

Enumerator
TT_NONE
TT_DELETED
TT_MAYBE_RAGGED
TT_MAYBE_ALIGNED
TT_CONFIRMED
TT_VLINE

Definition at line 60 of file blobbox.h.

             {
  TT_NONE,           // Not a tab.
  TT_DELETED,        // Not a tab after detailed analysis.
  TT_MAYBE_RAGGED,   // Initial designation of a tab-stop candidate.
  TT_MAYBE_ALIGNED,  // Initial designation of a tab-stop candidate.
  TT_CONFIRMED,      // Aligned with neighbours.
  TT_VLINE           // Detected as a vertical line.
};

Function Documentation

box_next()

TBOX box_next

(

BLOBNBOX_IT *

it

)

Definition at line 637 of file blobbox.cpp.

              {
  BLOBNBOX *blob;                //current blob
  TBOX result;                    //total box

  blob = it->data ();
  result = blob->bounding_box ();
  do {
    it->forward ();
    blob = it->data ();
    if (blob->cblob() == nullptr)
                                 //was pre-chopped
      result += blob->bounding_box ();
  }
                                 //until next real blob
  while ((blob->cblob() == nullptr) || blob->joined_to_prev());
  return result;
}

box_next_pre_chopped()

TBOX box_next_pre_chopped

(

BLOBNBOX_IT *

it

)

Definition at line 666 of file blobbox.cpp.

                          {
  BLOBNBOX *blob;                //current blob
  TBOX result;                    //total box

  blob = it->data ();
  result = blob->bounding_box ();
  do {
    it->forward ();
    blob = it->data ();
  }
                                 //until next real blob
  while (blob->joined_to_prev ());
  return result;
}

crotate_cblob()

C_BLOB* crotate_cblob	(	C_BLOB *	blob,
		FCOORD	rotation
	)

Definition at line 612 of file blobbox.cpp.

                       {
  C_OUTLINE_LIST out_list;       //output outlines
                                 //input outlines
  C_OUTLINE_IT in_it = blob->out_list ();
                                 //output outlines
  C_OUTLINE_IT out_it = &out_list;

  for (in_it.mark_cycle_pt (); !in_it.cycled_list (); in_it.forward ()) {
    out_it.add_after_then_move (new C_OUTLINE (in_it.data (), rotation));
  }
  return new C_BLOB (&out_list);
}

DirOtherWay()

BlobNeighbourDir DirOtherWay ( BlobNeighbourDir  dir )

inline

Definition at line 107 of file blobbox.h.

                                                          {
  return static_cast<BlobNeighbourDir>(dir ^ 2);
}

DominatesInMerge()

bool DominatesInMerge ( BlobTextFlowType  type1, BlobTextFlowType  type2  )

inline

Definition at line 130 of file blobbox.h.

                                                                             {
  // LEADER always loses.
  if (type1 == BTFT_LEADER) return false;
  if (type2 == BTFT_LEADER) return true;
  // With those out of the way, the ordering of the enum determines the result.
  return type1 >= type2;
}

find_cblob_hlimits()

void find_cblob_hlimits	(	C_BLOB *	blob,
		float	bottomy,
		float	topy,
		float &	xmin,
		float &	xymax
	)

Definition at line 577 of file blobbox.cpp.

                                     {
  int16_t stepindex;               //current point
  ICOORD pos;                    //current coords
  ICOORD vec;                    //rotated step
  C_OUTLINE *outline;            //current outline
                                 //outlines
  C_OUTLINE_IT out_it = blob->out_list ();

  xmin = (float) INT32_MAX;
  xmax = (float) -INT32_MAX;
  for (out_it.mark_cycle_pt (); !out_it.cycled_list (); out_it.forward ()) {
    outline = out_it.data ();
    pos = outline->start_pos (); //get coords
    for (stepindex = 0; stepindex < outline->pathlength (); stepindex++) {
                                 //inside
      if (pos.y () >= bottomy && pos.y () <= topy) {
        UpdateRange(pos.x(), &xmin, &xmax);
      }
      vec = outline->step (stepindex);
      pos += vec;                //move to next
    }
  }
}

find_cblob_limits()

void find_cblob_limits	(	C_BLOB *	blob,
		float	leftx,
		float	rightx,
		FCOORD	rotation,
		float &	ymin,
		float &	ymax
	)

Definition at line 500 of file blobbox.cpp.

                                    {
  int16_t stepindex;               //current point
  ICOORD pos;                    //current coords
  ICOORD vec;                    //rotated step
  C_OUTLINE *outline;            //current outline
                                 //outlines
  C_OUTLINE_IT out_it = blob->out_list ();

  ymin = (float) INT32_MAX;
  ymax = (float) -INT32_MAX;
  for (out_it.mark_cycle_pt (); !out_it.cycled_list (); out_it.forward ()) {
    outline = out_it.data ();
    pos = outline->start_pos (); //get coords
    pos.rotate (rotation);
    for (stepindex = 0; stepindex < outline->pathlength (); stepindex++) {
                                 //inside
      if (pos.x () >= leftx && pos.x () <= rightx) {
        UpdateRange(pos.y(), &ymin, &ymax);
      }
      vec = outline->step (stepindex);
      vec.rotate (rotation);
      pos += vec;                //move to next
    }
  }
}

find_cblob_vlimits()

void find_cblob_vlimits	(	C_BLOB *	blob,
		float	leftx,
		float	rightx,
		float &	ymin,
		float &	ymax
	)

Definition at line 540 of file blobbox.cpp.

                                     {
  int16_t stepindex;               //current point
  ICOORD pos;                    //current coords
  ICOORD vec;                    //rotated step
  C_OUTLINE *outline;            //current outline
                                 //outlines
  C_OUTLINE_IT out_it = blob->out_list ();

  ymin = (float) INT32_MAX;
  ymax = (float) -INT32_MAX;
  for (out_it.mark_cycle_pt (); !out_it.cycled_list (); out_it.forward ()) {
    outline = out_it.data ();
    pos = outline->start_pos (); //get coords
    for (stepindex = 0; stepindex < outline->pathlength (); stepindex++) {
                                 //inside
      if (pos.x () >= leftx && pos.x () <= rightx) {
        UpdateRange(pos.y(), &ymin, &ymax);
      }
      vec = outline->step (stepindex);
      pos += vec;                //move to next
    }
  }
}

plot_blob_list()

void plot_blob_list	(	ScrollView *	win,
		BLOBNBOX_LIST *	list,
		ScrollView::Color	body_colour,
		ScrollView::Color	child_colour
	)

Definition at line 1087 of file blobbox.cpp.

                                                  {  // colour of child
  BLOBNBOX_IT it = list;
  for (it.mark_cycle_pt(); !it.cycled_list(); it.forward()) {
    it.data()->plot(win, body_colour, child_colour);
  }
}

vertical_cblob_projection()

void vertical_cblob_projection	(	C_BLOB *	blob,
		STATS *	stats
	)

Definition at line 869 of file blobbox.cpp.

                                {
                                 //outlines of blob
  C_OUTLINE_IT out_it = blob->out_list ();

  for (out_it.mark_cycle_pt (); !out_it.cycled_list (); out_it.forward ()) {
    vertical_coutline_projection (out_it.data (), stats);
  }
}

vertical_coutline_projection()

void vertical_coutline_projection	(	C_OUTLINE *	outline,
		STATS *	stats
	)

Definition at line 889 of file blobbox.cpp.

                                   {
  ICOORD pos;                    //current point
  ICOORD step;                   //edge step
  int32_t length;                  //of outline
  int16_t stepindex;               //current step
  C_OUTLINE_IT out_it = outline->child ();

  pos = outline->start_pos ();
  length = outline->pathlength ();
  for (stepindex = 0; stepindex < length; stepindex++) {
    step = outline->step (stepindex);
    if (step.x () > 0) {
     stats->add (pos.x (), -pos.y ());
    } else if (step.x () < 0) {
      stats->add (pos.x () - 1, pos.y ());
    }
    pos += step;
  }

  for (out_it.mark_cycle_pt (); !out_it.cycled_list (); out_it.forward ()) {
    vertical_coutline_projection (out_it.data (), stats);
  }
}

Variable Documentation

textord_error_weight

double textord_error_weight = 3

"Weighting for error in believability"