#include <lm_state.h>

Inheritance diagram for tesseract::ViterbiStateEntry:

Public Member Functions
	ViterbiStateEntry (ViterbiStateEntry pe, BLOB_CHOICE b, float c, float ol, const LMConsistencyInfo &ci, const AssociateStats &as, LanguageModelFlagsType tcf, LanguageModelDawgInfo d, LanguageModelNgramInfo n, const char *debug_uch)

	~ViterbiStateEntry ()

bool	Consistent () const

bool	HasAlnumChoice (const UNICHARSET &unicharset)

void	Print (const char *msg) const

Public Member Functions inherited from tesseract::ELIST_LINK
	ELIST_LINK ()

	ELIST_LINK (const ELIST_LINK &)

void	operator= (const ELIST_LINK &)

Static Public Member Functions
static int	Compare (const void e1, const void e2)

Public Attributes
BLOB_CHOICE *	curr_b
	Pointers to BLOB_CHOICE and parent ViterbiStateEntry (not owned by this). More...

ViterbiStateEntry *	parent_vse

ViterbiStateEntry *	competing_vse

LanguageModelDawgInfo *	dawg_info

LanguageModelNgramInfo *	ngram_info

std::string *	debug_str

float	cost

float	ratings_sum
	sum of ratings of character on the path More...

float	min_certainty
	minimum certainty on the path More...

int	adapted
	number of BLOB_CHOICES from adapted templates More...

int	length
	number of characters on the path More...

float	outline_length
	length of the outline so far More...

LMConsistencyInfo	consistency_info
	path consistency info More...

AssociateStats	associate_stats
	character widths/gaps/seams More...

LanguageModelFlagsType	top_choice_flags

bool	updated
	set to true if the entry has just been created/updated More...

Detailed Description

Struct for storing the information about a path in the segmentation graph explored by Viterbi search.

Definition at line 92 of file lm_state.h.

Constructor & Destructor Documentation

◆ ViterbiStateEntry()

tesseract::ViterbiStateEntry::ViterbiStateEntry	(	ViterbiStateEntry *	pe,
		BLOB_CHOICE *	b,
		float	c,
		float	ol,
		const LMConsistencyInfo &	ci,
		const AssociateStats &	as,
		LanguageModelFlagsType	tcf,
		LanguageModelDawgInfo *	d,
		LanguageModelNgramInfo *	n,
		const char *	debug_uch
	)

inline

Definition at line 93 of file lm_state.h.

      : curr_b(b)
      , parent_vse(pe)
      , competing_vse(nullptr)
      , dawg_info(d)
      , ngram_info(n)
      , cost(c)
      , ratings_sum(b->rating())
      , min_certainty(b->certainty())
      , adapted(b->IsAdapted())
      , length(1)
      , outline_length(ol)
      , consistency_info(ci)
      , associate_stats(as)
      , top_choice_flags(tcf)
      , updated(true) {
    debug_str = (debug_uch == nullptr) ? nullptr : new std::string();
    if (pe != nullptr) {
      ratings_sum += pe->ratings_sum;
      if (pe->min_certainty < min_certainty) {
        min_certainty = pe->min_certainty;
      }
      adapted += pe->adapted;
      length += pe->length;
      outline_length += pe->outline_length;
      if (debug_uch != nullptr) {
        *debug_str += *(pe->debug_str);
      }
    }
    if (debug_str != nullptr && debug_uch != nullptr) {
      *debug_str += debug_uch;
    }
  }

◆ ~ViterbiStateEntry()

tesseract::ViterbiStateEntry::~ViterbiStateEntry ( )

inline

Definition at line 129 of file lm_state.h.

                       {
    delete dawg_info;
    delete ngram_info;
    delete debug_str;
  }

Member Function Documentation

◆ Compare()

static int tesseract::ViterbiStateEntry::Compare	(	const void *	e1,
		const void *	e2
	)

inlinestatic

Comparator function for sorting ViterbiStateEntry_LISTs in non-increasing order of costs.

Definition at line 136 of file lm_state.h.

                                                     {
    const ViterbiStateEntry *ve1 = *static_cast<const ViterbiStateEntry *const *>(e1);
    const ViterbiStateEntry *ve2 = *static_cast<const ViterbiStateEntry *const *>(e2);
    return (ve1->cost < ve2->cost) ? -1 : 1;
  }

◆ Consistent()

bool tesseract::ViterbiStateEntry::Consistent ( ) const

inline

Definition at line 141 of file lm_state.h.

                                 {
    if (dawg_info != nullptr && consistency_info.NumInconsistentCase() == 0) {
      return true;
    }
    return consistency_info.Consistent();
  }

◆ HasAlnumChoice()

bool tesseract::ViterbiStateEntry::HasAlnumChoice ( const UNICHARSET & unicharset )

inline

Returns true if this VSE has an alphanumeric character as its classifier result.

Definition at line 149 of file lm_state.h.

                                                    {
    if (curr_b == nullptr) {
      return false;
    }
    UNICHAR_ID unichar_id = curr_b->unichar_id();
    if (unicharset.get_isalpha(unichar_id) || unicharset.get_isdigit(unichar_id)) {
      return true;
    }
    return false;
  }

◆ Print()

void tesseract::ViterbiStateEntry::Print ( const char * msg ) const

Definition at line 24 of file lm_state.cpp.

                                                   {
  tprintf("%s ViterbiStateEntry", msg);
  if (updated) {
    tprintf("(NEW)");
  }
  if (this->debug_str != nullptr) {
    tprintf(" str=%s", this->debug_str->c_str());
  }
  tprintf(" with ratings_sum=%.4f length=%d cost=%.6f", this->ratings_sum, this->length,
          this->cost);
  if (this->top_choice_flags) {
    tprintf(" top_choice_flags=0x%x", this->top_choice_flags);
  }
  if (!this->Consistent()) {
    tprintf(" inconsistent=(punc %d case %d chartype %d script %d font %d)",
            this->consistency_info.NumInconsistentPunc(),
            this->consistency_info.NumInconsistentCase(),
            this->consistency_info.NumInconsistentChartype(),
            this->consistency_info.inconsistent_script, this->consistency_info.inconsistent_font);
  }
  if (this->dawg_info) {
    tprintf(" permuter=%d", this->dawg_info->permuter);
  }
  if (this->ngram_info) {
    tprintf(" ngram_cl_cost=%g context=%s ngram pruned=%d",
            this->ngram_info->ngram_and_classifier_cost, this->ngram_info->context.c_str(),
            this->ngram_info->pruned);
  }
  if (this->associate_stats.shape_cost > 0.0f) {
    tprintf(" shape_cost=%g", this->associate_stats.shape_cost);
  }
  tprintf(" %s", XHeightConsistencyEnumName[this->consistency_info.xht_decision]);
 
  tprintf("\n");
}

Member Data Documentation

◆ adapted

int tesseract::ViterbiStateEntry::adapted

number of BLOB_CHOICES from adapted templates

Definition at line 188 of file lm_state.h.

◆ associate_stats

AssociateStats tesseract::ViterbiStateEntry::associate_stats

character widths/gaps/seams

Definition at line 192 of file lm_state.h.

◆ competing_vse

ViterbiStateEntry* tesseract::ViterbiStateEntry::competing_vse

Pointer to a case-competing ViterbiStateEntry in the same list that represents a path ending in the same letter of the opposite case.

Definition at line 166 of file lm_state.h.

◆ consistency_info

LMConsistencyInfo tesseract::ViterbiStateEntry::consistency_info

path consistency info

Definition at line 191 of file lm_state.h.

◆ cost

float tesseract::ViterbiStateEntry::cost

The cost is an adjusted ratings sum, that is adjusted by all the language model components that use Viterbi search.

Definition at line 182 of file lm_state.h.

◆ curr_b

BLOB_CHOICE* tesseract::ViterbiStateEntry::curr_b

Pointers to BLOB_CHOICE and parent ViterbiStateEntry (not owned by this).

Definition at line 162 of file lm_state.h.

◆ dawg_info

LanguageModelDawgInfo* tesseract::ViterbiStateEntry::dawg_info

Extra information maintained by Dawg language model component (owned by ViterbiStateEntry).

Definition at line 170 of file lm_state.h.

◆ debug_str

std::string* tesseract::ViterbiStateEntry::debug_str

UTF8 string representing the path corresponding to this vse. Populated only in when language_model_debug_level > 0.

Definition at line 178 of file lm_state.h.

◆ length

int tesseract::ViterbiStateEntry::length

number of characters on the path

Definition at line 189 of file lm_state.h.

◆ min_certainty

float tesseract::ViterbiStateEntry::min_certainty

minimum certainty on the path

Definition at line 187 of file lm_state.h.

◆ ngram_info

LanguageModelNgramInfo* tesseract::ViterbiStateEntry::ngram_info

Extra information maintained by Ngram language model component (owned by ViterbiStateEntry).

Definition at line 174 of file lm_state.h.

◆ outline_length

float tesseract::ViterbiStateEntry::outline_length

length of the outline so far

Definition at line 190 of file lm_state.h.

◆ parent_vse

ViterbiStateEntry* tesseract::ViterbiStateEntry::parent_vse

Definition at line 163 of file lm_state.h.

◆ ratings_sum

float tesseract::ViterbiStateEntry::ratings_sum

sum of ratings of character on the path

Various information about the characters on the path represented by this ViterbiStateEntry.

Definition at line 186 of file lm_state.h.

◆ top_choice_flags

LanguageModelFlagsType tesseract::ViterbiStateEntry::top_choice_flags

Flags for marking the entry as a top choice path with the smallest rating or lower/upper case letters).

Definition at line 196 of file lm_state.h.

◆ updated

bool tesseract::ViterbiStateEntry::updated

set to true if the entry has just been created/updated

Definition at line 198 of file lm_state.h.

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

/media/home/debian/src/github/tesseract-ocr/tesseract/src/wordrec/lm_state.h
/media/home/debian/src/github/tesseract-ocr/tesseract/src/wordrec/lm_state.cpp

Public Member Functions

Static Public Member Functions

Public Attributes

Detailed Description

Constructor & Destructor Documentation

◆ ViterbiStateEntry()

◆ ~ViterbiStateEntry()

Member Function Documentation

◆ Compare()

◆ Consistent()

◆ HasAlnumChoice()

◆ Print()

Member Data Documentation

◆ adapted

◆ associate_stats

◆ competing_vse

◆ consistency_info

◆ cost

◆ curr_b

◆ dawg_info

◆ debug_str

◆ length

◆ min_certainty

◆ ngram_info

◆ outline_length

◆ parent_vse

◆ ratings_sum

◆ top_choice_flags

◆ updated