#include <strokewidth.h>

Inheritance diagram for tesseract::StrokeWidth:

Public Member Functions
	StrokeWidth (int gridsize, const ICOORD &bleft, const ICOORD &tright)

	~StrokeWidth () override

void	SetNeighboursOnMediumBlobs (TO_BLOCK *block)

void	FindTextlineDirectionAndFixBrokenCJK (PageSegMode pageseg_mode, bool cjk_merge, TO_BLOCK *input_block)

bool	TestVerticalTextDirection (double find_vertical_text_ratio, TO_BLOCK block, BLOBNBOX_CLIST osd_blobs)

void	CorrectForRotation (const FCOORD &rerotation, ColPartitionGrid *part_grid)

void	FindLeaderPartitions (TO_BLOCK block, ColPartitionGrid part_grid)

void	RemoveLineResidue (ColPartition_LIST *big_part_list)

void	GradeBlobsIntoPartitions (PageSegMode pageseg_mode, const FCOORD &rerotation, TO_BLOCK block, Image nontext_pix, const DENORM denorm, bool cjk_script, TextlineProjection projection, BLOBNBOX_LIST diacritic_blobs, ColPartitionGrid part_grid, ColPartition_LIST big_parts)

void	HandleClick (int x, int y) override

Public Member Functions inherited from tesseract::BlobGrid
	BlobGrid (int gridsize, const ICOORD &bleft, const ICOORD &tright)

	~BlobGrid () override

void	InsertBlobList (BLOBNBOX_LIST *blobs)

Public Member Functions inherited from tesseract::BBGrid< BLOBNBOX, BLOBNBOX_CLIST, BLOBNBOX_C_IT >
	BBGrid ()

	BBGrid (int gridsize, const ICOORD &bleft, const ICOORD &tright)

	~BBGrid () override

void	Init (int gridsize, const ICOORD &bleft, const ICOORD &tright)

void	Clear ()

void	ClearGridData (void(free_method)(BLOBNBOX ))

void	InsertBBox (bool h_spread, bool v_spread, BLOBNBOX *bbox)

void	InsertPixPtBBox (int left, int bottom, Image pix, BLOBNBOX *bbox)

void	RemoveBBox (BLOBNBOX *bbox)

bool	RectangleEmpty (const TBOX &rect)

IntGrid *	CountCellElements ()

ScrollView *	MakeWindow (int x, int y, const char *window_name)

void	DisplayBoxes (ScrollView *window)

void	AssertNoDuplicates ()

virtual void	HandleClick (int x, int y)

Public Member Functions inherited from tesseract::GridBase
	GridBase ()=default

	GridBase (int gridsize, const ICOORD &bleft, const ICOORD &tright)

virtual	~GridBase ()

void	Init (int gridsize, const ICOORD &bleft, const ICOORD &tright)

int	gridsize () const

int	gridwidth () const

int	gridheight () const

const ICOORD &	bleft () const

const ICOORD &	tright () const

void	GridCoords (int x, int y, int grid_x, int grid_y) const

void	ClipGridCoords (int x, int y) const

Additional Inherited Members
Protected Attributes inherited from tesseract::BBGrid< BLOBNBOX, BLOBNBOX_CLIST, BLOBNBOX_C_IT >
BLOBNBOX_CLIST *	grid_

Protected Attributes inherited from tesseract::GridBase
int	gridsize_

int	gridwidth_

int	gridheight_

int	gridbuckets_

ICOORD	bleft_

ICOORD	tright_

Detailed Description

The StrokeWidth class holds all the normal and large blobs. It is used to find good large blobs and move them to the normal blobs by virtue of having a reasonable strokewidth compatible neighbour.

Definition at line 53 of file strokewidth.h.

Constructor & Destructor Documentation

◆ StrokeWidth()

tesseract::StrokeWidth::StrokeWidth	(	int	gridsize,
		const ICOORD &	bleft,
		const ICOORD &	tright
	)

Definition at line 114 of file strokewidth.cpp.

    : BlobGrid(gridsize, bleft, tright)
    , nontext_map_(nullptr)
    , projection_(nullptr)
    , denorm_(nullptr)
    , grid_box_(bleft, tright)
    , rerotation_(1.0f, 0.0f) {
}

◆ ~StrokeWidth()

tesseract::StrokeWidth::~StrokeWidth ( )

override

Definition at line 123 of file strokewidth.cpp.

                          {
#ifndef GRAPHICS_DISABLED
  if (widths_win_ != nullptr) {
    widths_win_->AwaitEvent(SVET_DESTROY);
    if (textord_tabfind_only_strokewidths) {
      exit(0);
    }
    delete widths_win_;
  }
  delete leaders_win_;
  delete initial_widths_win_;
  delete chains_win_;
  delete textlines_win_;
  delete smoothed_win_;
  delete diacritics_win_;
#endif
}

Member Function Documentation

◆ CorrectForRotation()

void tesseract::StrokeWidth::CorrectForRotation	(	const FCOORD &	rerotation,
		ColPartitionGrid *	part_grid
	)

Definition at line 252 of file strokewidth.cpp.

                                                                                        {
  Init(part_grid->gridsize(), part_grid->bleft(), part_grid->tright());
  grid_box_ = TBOX(bleft(), tright());
  rerotation_.set_x(rotation.x());
  rerotation_.set_y(-rotation.y());
}

◆ FindLeaderPartitions()

void tesseract::StrokeWidth::FindLeaderPartitions	(	TO_BLOCK *	block,
		ColPartitionGrid *	part_grid
	)

Definition at line 260 of file strokewidth.cpp.

                                                                                   {
  Clear();
  // Find and isolate leaders in the noise list.
  ColPartition_LIST leader_parts;
  FindLeadersAndMarkNoise(block, &leader_parts);
  // Setup the strokewidth grid with the block's remaining (non-noise) blobs.
  InsertBlobList(&block->blobs);
  // Mark blobs that have leader neighbours.
  for (ColPartition_IT it(&leader_parts); !it.empty(); it.forward()) {
    ColPartition *part = it.extract();
    part->ClaimBoxes();
    MarkLeaderNeighbours(part, LR_LEFT);
    MarkLeaderNeighbours(part, LR_RIGHT);
    part_grid->InsertBBox(true, true, part);
  }
}

◆ FindTextlineDirectionAndFixBrokenCJK()

void tesseract::StrokeWidth::FindTextlineDirectionAndFixBrokenCJK	(	PageSegMode	pageseg_mode,
		bool	cjk_merge,
		TO_BLOCK *	input_block
	)

Definition at line 158 of file strokewidth.cpp.

                                                                              {
  // Setup the grid with the remaining (non-noise) blobs.
  InsertBlobs(input_block);
  // Repair broken CJK characters if needed.
  while (cjk_merge && FixBrokenCJK(input_block)) {
  }
  // Grade blobs by inspection of neighbours.
  FindTextlineFlowDirection(pageseg_mode, false);
  // Clear the grid ready for rotation or leader finding.
  Clear();
}

◆ GradeBlobsIntoPartitions()

void tesseract::StrokeWidth::GradeBlobsIntoPartitions	(	PageSegMode	pageseg_mode,
		const FCOORD &	rerotation,
		TO_BLOCK *	block,
		Image	nontext_pix,
		const DENORM *	denorm,
		bool	cjk_script,
		TextlineProjection *	projection,
		BLOBNBOX_LIST *	diacritic_blobs,
		ColPartitionGrid *	part_grid,
		ColPartition_LIST *	big_parts
	)

Definition at line 344 of file strokewidth.cpp.

                                                                         {
  nontext_map_ = nontext_pix;
  projection_ = projection;
  denorm_ = denorm;
  // Clear and re Insert to take advantage of the tab stops in the blobs.
  Clear();
  // Setup the strokewidth grid with the remaining non-noise, non-leader blobs.
  InsertBlobs(block);
 
  // Run FixBrokenCJK() again if the page is CJK.
  if (cjk_script) {
    FixBrokenCJK(block);
  }
  FindTextlineFlowDirection(pageseg_mode, false);
  projection_->ConstructProjection(block, rerotation, nontext_map_);
#ifndef GRAPHICS_DISABLED
  if (textord_tabfind_show_strokewidths) {
    ScrollView *line_blobs_win = MakeWindow(0, 0, "Initial textline Blobs");
    projection_->PlotGradedBlobs(&block->blobs, line_blobs_win);
    projection_->PlotGradedBlobs(&block->small_blobs, line_blobs_win);
  }
#endif
  projection_->MoveNonTextlineBlobs(&block->blobs, &block->noise_blobs);
  projection_->MoveNonTextlineBlobs(&block->small_blobs, &block->noise_blobs);
  // Clear and re Insert to take advantage of the removed diacritics.
  Clear();
  InsertBlobs(block);
  FCOORD skew;
  FindTextlineFlowDirection(pageseg_mode, true);
  PartitionFindResult r = FindInitialPartitions(pageseg_mode, rerotation, true, block,
                                                diacritic_blobs, part_grid, big_parts, &skew);
  if (r == PFR_NOISE) {
    tprintf("Detected %d diacritics\n", diacritic_blobs->length());
    // Noise was found, and removed.
    Clear();
    InsertBlobs(block);
    FindTextlineFlowDirection(pageseg_mode, true);
    r = FindInitialPartitions(pageseg_mode, rerotation, false, block, diacritic_blobs, part_grid,
                              big_parts, &skew);
  }
  nontext_map_ = nullptr;
  projection_ = nullptr;
  denorm_ = nullptr;
}

◆ HandleClick()

void tesseract::StrokeWidth::HandleClick	(	int	x,
		int	y
	)

overridevirtual

Handles a click event in a display window.

Reimplemented from tesseract::BBGrid< BLOBNBOX, BLOBNBOX_CLIST, BLOBNBOX_C_IT >.

Definition at line 403 of file strokewidth.cpp.

                                          {
  BBGrid<BLOBNBOX, BLOBNBOX_CLIST, BLOBNBOX_C_IT>::HandleClick(x, y);
  // Run a radial search for blobs that overlap.
  BlobGridSearch radsearch(this);
  radsearch.StartRadSearch(x, y, 1);
  BLOBNBOX *neighbour;
  FCOORD click(static_cast<float>(x), static_cast<float>(y));
  while ((neighbour = radsearch.NextRadSearch()) != nullptr) {
    TBOX nbox = neighbour->bounding_box();
    if (nbox.contains(click) && neighbour->cblob() != nullptr) {
      PrintBoxWidths(neighbour);
      if (neighbour->neighbour(BND_LEFT) != nullptr) {
        PrintBoxWidths(neighbour->neighbour(BND_LEFT));
      }
      if (neighbour->neighbour(BND_RIGHT) != nullptr) {
        PrintBoxWidths(neighbour->neighbour(BND_RIGHT));
      }
      if (neighbour->neighbour(BND_ABOVE) != nullptr) {
        PrintBoxWidths(neighbour->neighbour(BND_ABOVE));
      }
      if (neighbour->neighbour(BND_BELOW) != nullptr) {
        PrintBoxWidths(neighbour->neighbour(BND_BELOW));
      }
      int gaps[BND_COUNT];
      neighbour->NeighbourGaps(gaps);
      tprintf(
          "Left gap=%d, right=%d, above=%d, below=%d, horz=%d, vert=%d\n"
          "Good=    %d        %d        %d        %d\n",
          gaps[BND_LEFT], gaps[BND_RIGHT], gaps[BND_ABOVE], gaps[BND_BELOW],
          neighbour->horz_possible(), neighbour->vert_possible(),
          neighbour->good_stroke_neighbour(BND_LEFT), neighbour->good_stroke_neighbour(BND_RIGHT),
          neighbour->good_stroke_neighbour(BND_ABOVE), neighbour->good_stroke_neighbour(BND_BELOW));
      break;
    }
  }
}

◆ RemoveLineResidue()

void tesseract::StrokeWidth::RemoveLineResidue ( ColPartition_LIST * big_part_list )

Definition at line 279 of file strokewidth.cpp.

                                                                    {
  BlobGridSearch gsearch(this);
  BLOBNBOX *bbox;
  // For every vertical line-like bbox in the grid, search its neighbours
  // to find the tallest, and if the original box is taller by sufficient
  // margin, then call it line residue and delete it.
  gsearch.StartFullSearch();
  while ((bbox = gsearch.NextFullSearch()) != nullptr) {
    TBOX box = bbox->bounding_box();
    if (box.height() < box.width() * kLineResidueAspectRatio) {
      continue;
    }
    // Set up a rectangle search around the blob to find the size of its
    // neighbours.
    int padding = box.height() * kLineResiduePadRatio;
    TBOX search_box = box;
    search_box.pad(padding, padding);
    bool debug = AlignedBlob::WithinTestRegion(2, box.left(), box.bottom());
    // Find the largest object in the search box not equal to bbox.
    BlobGridSearch rsearch(this);
    int max_height = 0;
    BLOBNBOX *n;
    rsearch.StartRectSearch(search_box);
    while ((n = rsearch.NextRectSearch()) != nullptr) {
      if (n == bbox) {
        continue;
      }
      TBOX nbox = n->bounding_box();
      if (nbox.height() > max_height) {
        max_height = nbox.height();
      }
    }
    if (debug) {
      tprintf("Max neighbour size=%d for candidate line box at:", max_height);
      box.print();
    }
    if (max_height * kLineResidueSizeRatio < box.height()) {
#ifndef GRAPHICS_DISABLED
      if (leaders_win_ != nullptr) {
        // We are debugging, so display deleted in pink blobs in the same
        // window that we use to display leader detection.
        leaders_win_->Pen(ScrollView::PINK);
        leaders_win_->Rectangle(box.left(), box.bottom(), box.right(), box.top());
      }
#endif // !GRAPHICS_DISABLED
      ColPartition::MakeBigPartition(bbox, big_part_list);
    }
  }
}

◆ SetNeighboursOnMediumBlobs()

void tesseract::StrokeWidth::SetNeighboursOnMediumBlobs ( TO_BLOCK * block )

Definition at line 144 of file strokewidth.cpp.

                                                            {
  // Run a preliminary strokewidth neighbour detection on the medium blobs.
  InsertBlobList(&block->blobs);
  BLOBNBOX_IT blob_it(&block->blobs);
  for (blob_it.mark_cycle_pt(); !blob_it.cycled_list(); blob_it.forward()) {
    SetNeighbours(false, false, blob_it.data());
  }
  Clear();
}

◆ TestVerticalTextDirection()

bool tesseract::StrokeWidth::TestVerticalTextDirection	(	double	find_vertical_text_ratio,
		TO_BLOCK *	block,
		BLOBNBOX_CLIST *	osd_blobs
	)

Definition at line 211 of file strokewidth.cpp.

                                                                       {
  int vertical_boxes = 0;
  int horizontal_boxes = 0;
  // Count vertical normal and large blobs.
  BLOBNBOX_CLIST vertical_blobs;
  BLOBNBOX_CLIST horizontal_blobs;
  BLOBNBOX_CLIST nondescript_blobs;
  CollectHorizVertBlobs(&block->blobs, &vertical_boxes, &horizontal_boxes, &vertical_blobs,
                        &horizontal_blobs, &nondescript_blobs);
  CollectHorizVertBlobs(&block->large_blobs, &vertical_boxes, &horizontal_boxes, &vertical_blobs,
                        &horizontal_blobs, &nondescript_blobs);
  if (textord_debug_tabfind) {
    tprintf("TextDir hbox=%d vs vbox=%d, %dH, %dV, %dN osd blobs\n", horizontal_boxes,
            vertical_boxes, horizontal_blobs.length(), vertical_blobs.length(),
            nondescript_blobs.length());
  }
  if (osd_blobs != nullptr && vertical_boxes == 0 && horizontal_boxes == 0) {
    // Only nondescript blobs available, so return those.
    BLOBNBOX_C_IT osd_it(osd_blobs);
    osd_it.add_list_after(&nondescript_blobs);
    return false;
  }
  int min_vert_boxes =
      static_cast<int>((vertical_boxes + horizontal_boxes) * find_vertical_text_ratio);
  if (vertical_boxes >= min_vert_boxes) {
    if (osd_blobs != nullptr) {
      BLOBNBOX_C_IT osd_it(osd_blobs);
      osd_it.add_list_after(&vertical_blobs);
    }
    return true;
  } else {
    if (osd_blobs != nullptr) {
      BLOBNBOX_C_IT osd_it(osd_blobs);
      osd_it.add_list_after(&horizontal_blobs);
    }
    return false;
  }
}

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

/media/home/debian/src/github/tesseract-ocr/tesseract/src/textord/strokewidth.h
/media/home/debian/src/github/tesseract-ocr/tesseract/src/textord/strokewidth.cpp

Public Member Functions

Additional Inherited Members