tesseract v5.3.3.20231005
Public Member Functions | Static Public Member Functions | Public Attributes | List of all members
tesseract::TWERD Struct Reference

#include <blobs.h>

Public Member Functions

 TWERD ()
 
 TWERD (const TWERD &src)
 
 ~TWERD ()
 
TWERD & operator= (const TWERD &src)
 
void BLNormalize (const BLOCK *block, const ROW *row, Image pix, bool inverse, float x_height, float baseline_shift, bool numeric_mode, tesseract::OcrEngineMode hint, const TBOX *norm_box, DENORM *word_denorm)
 
void CopyFrom (const TWERD &src)
 
void Clear ()
 
void ComputeBoundingBoxes ()
 
unsigned NumBlobs () const
 
TBOX bounding_box () const
 
void MergeBlobs (unsigned start, unsigned end)
 
void plot (ScrollView *window)
 

Static Public Member Functions

static TWERD * PolygonalCopy (bool allow_detailed_fx, WERD *src)
 

Public Attributes

std::vector< TBLOB * > blobs
 
bool latin_script
 

Detailed Description

Definition at line 421 of file blobs.h.

Constructor & Destructor Documentation

◆ TWERD() [1/2]

tesseract::TWERD::TWERD ( )
inline

Definition at line 422 of file blobs.h.

422: latin_script(false) {}
tesseract::TWERD::latin_script
bool latin_script
Definition: blobs.h:463

◆ TWERD() [2/2]

tesseract::TWERD::TWERD ( const TWERD &  src)
inline

Definition at line 423 of file blobs.h.

423 {
424 CopyFrom(src);
425 }
tesseract::TWERD::CopyFrom
void CopyFrom(const TWERD &src)
Definition: blobs.cpp:839

◆ ~TWERD()

tesseract::TWERD::~TWERD ( )
inline

Definition at line 426 of file blobs.h.

426 {
427 Clear();
428 }
tesseract::TWERD::Clear
void Clear()
Definition: blobs.cpp:849

Member Function Documentation

◆ BLNormalize()

void tesseract::TWERD::BLNormalize ( const BLOCK *  block,
const ROW *  row,
Image  pix,
bool  inverse,
float  x_height,
float  baseline_shift,
bool  numeric_mode,
tesseract::OcrEngineMode  hint,
const TBOX *  norm_box,
DENORM *  word_denorm 
)

Definition at line 792 of file blobs.cpp.

794 {
795 TBOX word_box = bounding_box();
796 if (norm_box != nullptr) {
797 word_box = *norm_box;
798 }
799 float word_middle = (word_box.left() + word_box.right()) / 2.0f;
800 float input_y_offset = 0.0f;
801 auto final_y_offset = static_cast<float>(kBlnBaselineOffset);
802 float scale = kBlnXHeight / x_height;
803 if (row == nullptr) {
804 word_middle = word_box.left();
805 input_y_offset = word_box.bottom();
806 final_y_offset = 0.0f;
807 } else {
808 input_y_offset = row->base_line(word_middle) + baseline_shift;
809 }
810 for (auto blob : blobs) {
811 TBOX blob_box = blob->bounding_box();
812 float mid_x = (blob_box.left() + blob_box.right()) / 2.0f;
813 float baseline = input_y_offset;
814 float blob_scale = scale;
815 if (numeric_mode) {
816 baseline = blob_box.bottom();
817 blob_scale = ClipToRange(kBlnXHeight * 4.0f / (3 * blob_box.height()), scale, scale * 1.5f);
818 } else if (row != nullptr) {
819 baseline = row->base_line(mid_x) + baseline_shift;
820 }
821 // The image will be 8-bit grey if the input was grey or color. Note that in
822 // a grey image 0 is black and 255 is white. If the input was binary, then
823 // the pix will be binary and 0 is white, with 1 being black.
824 // To tell the difference pixGetDepth() will return 8 or 1.
825 // The inverse flag will be true iff the word has been determined to be
826 // white on black, and is independent of whether the pix is 8 bit or 1 bit.
827 blob->Normalize(block, nullptr, nullptr, word_middle, baseline, blob_scale, blob_scale, 0.0f,
828 final_y_offset, inverse, pix);
829 }
830 if (word_denorm != nullptr) {
831 word_denorm->SetupNormalization(block, nullptr, nullptr, word_middle, input_y_offset, scale,
832 scale, 0.0f, final_y_offset);
833 word_denorm->set_inverse(inverse);
834 word_denorm->set_pix(pix);
835 }
836}
TBOX
@ TBOX
Definition: cleanapi_test.cc:19
tesseract::baseline
@ baseline
Definition: mfoutline.h:53
tesseract::kBlnXHeight
const int kBlnXHeight
Definition: normalis.h:33
tesseract::ClipToRange
T ClipToRange(const T &x, const T &lower_bound, const T &upper_bound)
Definition: helpers.h:105
tesseract::kBlnBaselineOffset
const int kBlnBaselineOffset
Definition: normalis.h:34
tesseract::TWERD::bounding_box
TBOX bounding_box() const
Definition: blobs.cpp:863
tesseract::TWERD::blobs
std::vector< TBLOB * > blobs
Definition: blobs.h:462

◆ bounding_box()

TBOX tesseract::TWERD::bounding_box ( ) const

Definition at line 863 of file blobs.cpp.

863 {
864 TBOX result;
865 for (auto blob : blobs) {
866 TBOX box = blob->bounding_box();
867 result += box;
868 }
869 return result;
870}

◆ Clear()

void tesseract::TWERD::Clear ( )

Definition at line 849 of file blobs.cpp.

849 {
850 for (auto blob : blobs) {
851 delete blob;
852 }
853 blobs.clear();
854}

◆ ComputeBoundingBoxes()

void tesseract::TWERD::ComputeBoundingBoxes ( )

Definition at line 857 of file blobs.cpp.

857 {
858 for (auto &blob : blobs) {
859 blob->ComputeBoundingBoxes();
860 }
861}

◆ CopyFrom()

void tesseract::TWERD::CopyFrom ( const TWERD &  src)

Definition at line 839 of file blobs.cpp.

839 {
840 Clear();
841 latin_script = src.latin_script;
842 for (auto blob : src.blobs) {
843 auto *new_blob = new TBLOB(*blob);
844 blobs.push_back(new_blob);
845 }
846}

◆ MergeBlobs()

void tesseract::TWERD::MergeBlobs ( unsigned  start,
unsigned  end 
)

Definition at line 874 of file blobs.cpp.

874 {
875 if (end > blobs.size()) {
876 end = blobs.size();
877 }
878 if (start >= end) {
879 return; // Nothing to do.
880 }
881 TESSLINE *outline = blobs[start]->outlines;
882 for (auto i = start + 1; i < end; ++i) {
883 TBLOB *next_blob = blobs[i];
884 // Take the outlines from the next blob.
885 if (outline == nullptr) {
886 blobs[start]->outlines = next_blob->outlines;
887 outline = blobs[start]->outlines;
888 } else {
889 while (outline->next != nullptr) {
890 outline = outline->next;
891 }
892 outline->next = next_blob->outlines;
893 next_blob->outlines = nullptr;
894 }
895 // Delete the next blob and move on.
896 delete next_blob;
897 blobs[i] = nullptr;
898 }
899 // Remove dead blobs from the vector.
900 // TODO: optimize.
901 for (auto i = start + 1; i < end && start + 1 < blobs.size(); ++i) {
902 blobs.erase(blobs.begin() + start + 1);
903 }
904}

◆ NumBlobs()

unsigned tesseract::TWERD::NumBlobs ( ) const
inline

Definition at line 449 of file blobs.h.

449 {
450 return blobs.size();
451 }

◆ operator=()

TWERD & tesseract::TWERD::operator= ( const TWERD &  src)
inline

Definition at line 429 of file blobs.h.

429 {
430 CopyFrom(src);
431 return *this;
432 }

◆ plot()

void tesseract::TWERD::plot ( ScrollView *  window)

Definition at line 907 of file blobs.cpp.

907 {
908 ScrollView::Color color = WERD::NextColor(ScrollView::BLACK);
909 for (auto &blob : blobs) {
910 blob->plot(window, color, ScrollView::BROWN);
911 color = WERD::NextColor(color);
912 }
913}
tesseract::WERD::NextColor
static ScrollView::Color NextColor(ScrollView::Color colour)
Definition: werd.cpp:298

◆ PolygonalCopy()

TWERD * tesseract::TWERD::PolygonalCopy ( bool  allow_detailed_fx,
WERD *  src 
)
static

Definition at line 778 of file blobs.cpp.

778 {
779 auto *tessword = new TWERD;
780 tessword->latin_script = src->flag(W_SCRIPT_IS_LATIN);
781 C_BLOB_IT b_it(src->cblob_list());
782 for (b_it.mark_cycle_pt(); !b_it.cycled_list(); b_it.forward()) {
783 C_BLOB *blob = b_it.data();
784 TBLOB *tblob = TBLOB::PolygonalCopy(allow_detailed_fx, blob);
785 tessword->blobs.push_back(tblob);
786 }
787 return tessword;
788}
tesseract::W_SCRIPT_IS_LATIN
@ W_SCRIPT_IS_LATIN
Special case latin for y. splitting.
Definition: werd.h:38
tesseract::TBLOB::PolygonalCopy
static TBLOB * PolygonalCopy(bool allow_detailed_fx, C_BLOB *src)
Definition: blobs.cpp:335

Member Data Documentation

◆ blobs

std::vector<TBLOB *> tesseract::TWERD::blobs

Definition at line 462 of file blobs.h.

◆ latin_script

bool tesseract::TWERD::latin_script

Definition at line 463 of file blobs.h.

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