tesseract::IntegerMatcher Class Reference

#include <intmatcher.h>

Public Member Functions
	IntegerMatcher (tesseract::IntParam *classify_debug_level)
void	Match (INT_CLASS_STRUCT *ClassTemplate, BIT_VECTOR ProtoMask, BIT_VECTOR ConfigMask, int16_t NumFeatures, const INT_FEATURE_STRUCT *Features, tesseract::UnicharRating *Result, int AdaptFeatureThreshold, int Debug, bool SeparateDebugWindows)
float	ApplyCNCorrection (float rating, int blob_length, int normalization_factor, int matcher_multiplier)
int	FindGoodProtos (INT_CLASS_STRUCT *ClassTemplate, BIT_VECTOR ProtoMask, BIT_VECTOR ConfigMask, int16_t NumFeatures, INT_FEATURE_ARRAY Features, PROTO_ID *ProtoArray, int AdaptProtoThreshold, int Debug)
int	FindBadFeatures (INT_CLASS_STRUCT *ClassTemplate, BIT_VECTOR ProtoMask, BIT_VECTOR ConfigMask, int16_t NumFeatures, INT_FEATURE_ARRAY Features, FEATURE_ID *FeatureArray, int AdaptFeatureThreshold, int Debug)

Static Public Attributes
static const int	kIntThetaFudge = 128
static const int	kEvidenceTableBits = 9
static const int	kIntEvidenceTruncBits = 14
static const float	kSEExponentialMultiplier = 0.0f
static const float	kSimilarityCenter = 0.0075f

Detailed Description

Definition at line 60 of file intmatcher.h.

Constructor & Destructor Documentation

IntegerMatcher()

tesseract::IntegerMatcher::IntegerMatcher

(

tesseract::IntParam *

classify_debug_level

)

Definition at line 668 of file intmatcher.cpp.

    : classify_debug_level_(classify_debug_level) {
  /* Initialize table for evidence to similarity lookup */
  for (int i = 0; i < SE_TABLE_SIZE; i++) {
    uint32_t IntSimilarity = i << (27 - SE_TABLE_BITS);
    double Similarity = (static_cast<double>(IntSimilarity)) / 65536.0 / 65536.0;
    double evidence = Similarity / kSimilarityCenter;
    evidence = 255.0 / (evidence * evidence + 1.0);
 
    if (kSEExponentialMultiplier > 0.0) {
      double scale =
          1.0 - std::exp(-kSEExponentialMultiplier) *
                    exp(kSEExponentialMultiplier * (static_cast<double>(i) / SE_TABLE_SIZE));
      evidence *= ClipToRange(scale, 0.0, 1.0);
    }
 
    similarity_evidence_table_[i] = static_cast<uint8_t>(evidence + 0.5);
  }
 
  /* Initialize evidence computation variables */
  evidence_table_mask_ = ((1 << kEvidenceTableBits) - 1) << (9 - kEvidenceTableBits);
  mult_trunc_shift_bits_ = (14 - kIntEvidenceTruncBits);
  table_trunc_shift_bits_ = (27 - SE_TABLE_BITS - (mult_trunc_shift_bits_ << 1));
  evidence_mult_mask_ = ((1 << kIntEvidenceTruncBits) - 1);
}

Member Function Documentation

ApplyCNCorrection()

float tesseract::IntegerMatcher::ApplyCNCorrection	(	float	rating,
		int	blob_length,
		int	normalization_factor,
		int	matcher_multiplier
	)

Applies the CN normalization factor to the given rating and returns the modified rating.

Definition at line 1156 of file intmatcher.cpp.

                                                                {
  int divisor = blob_length + matcher_multiplier;
  return divisor == 0
             ? 1.0f
             : (rating * blob_length + matcher_multiplier * normalization_factor / 256.0f) /
                   divisor;
}

FindBadFeatures()

int tesseract::IntegerMatcher::FindBadFeatures	(	INT_CLASS_STRUCT *	ClassTemplate,
		BIT_VECTOR	ProtoMask,
		BIT_VECTOR	ConfigMask,
		int16_t	NumFeatures,
		INT_FEATURE_ARRAY	Features,
		FEATURE_ID *	FeatureArray,
		int	AdaptFeatureThreshold,
		int	Debug
	)

FindBadFeatures finds all features with maximum feature-evidence < AdaptFeatureThresh. The list is ordered by increasing feature number.

Parameters

ClassTemplate	Prototypes & tables for a class
ProtoMask	AND Mask for proto word
ConfigMask	AND Mask for config word
NumFeatures	Number of features in blob
Features	Array of features
FeatureArray	Array of bad features
AdaptFeatureThreshold	Threshold for bad features
Debug	Debugger flag: 1=debugger on

Returns

Number of bad features in FeatureArray.

Definition at line 619 of file intmatcher.cpp.

                                                                          {
  auto *tables = new ScratchEvidence();
  int NumBadFeatures = 0;
 
  /* DEBUG opening heading */
  if (MatchDebuggingOn(Debug)) {
    tprintf("Find Bad Features -------------------------------------------\n");
  }
 
  tables->Clear(ClassTemplate);
 
  for (int Feature = 0; Feature < NumFeatures; Feature++) {
    UpdateTablesForFeature(ClassTemplate, ProtoMask, ConfigMask, Feature, &Features[Feature],
                           tables, Debug);
 
    /* Find Best Evidence for Current Feature */
    int best = 0;
    assert(ClassTemplate->NumConfigs < MAX_NUM_CONFIGS);
    for (int i = 0; i < MAX_NUM_CONFIGS && i < ClassTemplate->NumConfigs; i++) {
      if (tables->feature_evidence_[i] > best) {
        best = tables->feature_evidence_[i];
      }
    }
 
    /* Find Bad Features */
    if (best < AdaptFeatureThreshold) {
      *FeatureArray = Feature;
      FeatureArray++;
      NumBadFeatures++;
    }
  }
 
#ifndef GRAPHICS_DISABLED
  if (PrintProtoMatchesOn(Debug) || PrintMatchSummaryOn(Debug)) {
    DebugFeatureProtoError(ClassTemplate, ProtoMask, ConfigMask, *tables, NumFeatures, Debug);
  }
#endif
 
  if (MatchDebuggingOn(Debug)) {
    tprintf("Match Complete --------------------------------------------\n");
  }
 
  delete tables;
  return NumBadFeatures;
}

FindGoodProtos()

int tesseract::IntegerMatcher::FindGoodProtos	(	INT_CLASS_STRUCT *	ClassTemplate,
		BIT_VECTOR	ProtoMask,
		BIT_VECTOR	ConfigMask,
		int16_t	NumFeatures,
		INT_FEATURE_ARRAY	Features,
		PROTO_ID *	ProtoArray,
		int	AdaptProtoThreshold,
		int	Debug
	)

FindGoodProtos finds all protos whose normalized proto-evidence exceed AdaptProtoThreshold. The list is ordered by increasing proto id number.

Globals:

• local_matcher_multiplier_ Normalization factor multiplier param ClassTemplate Prototypes & tables for a class param ProtoMask AND Mask for proto word param ConfigMask AND Mask for config word param NumFeatures Number of features in blob param Features Array of features param ProtoArray Array of good protos param AdaptProtoThreshold Threshold for good protos param Debug Debugger flag: 1=debugger on

  Returns

  Number of good protos in ProtoArray.

Definition at line 555 of file intmatcher.cpp.

                                                                       {
  auto *tables = new ScratchEvidence();
  int NumGoodProtos = 0;
 
  /* DEBUG opening heading */
  if (MatchDebuggingOn(Debug)) {
    tprintf("Find Good Protos -------------------------------------------\n");
  }
 
  tables->Clear(ClassTemplate);
 
  for (int Feature = 0; Feature < NumFeatures; Feature++) {
    UpdateTablesForFeature(ClassTemplate, ProtoMask, ConfigMask, Feature, &(Features[Feature]),
                           tables, Debug);
  }
 
#ifndef GRAPHICS_DISABLED
  if (PrintProtoMatchesOn(Debug) || PrintMatchSummaryOn(Debug)) {
    DebugFeatureProtoError(ClassTemplate, ProtoMask, ConfigMask, *tables, NumFeatures, Debug);
  }
#endif
 
  /* Average Proto Evidences & Find Good Protos */
  for (int proto = 0; proto < ClassTemplate->NumProtos; proto++) {
    /* Compute Average for Actual Proto */
    int Temp = 0;
    for (uint8_t i = 0; i < MAX_PROTO_INDEX && i < ClassTemplate->ProtoLengths[proto]; i++) {
      Temp += tables->proto_evidence_[proto][i];
    }
 
    Temp /= ClassTemplate->ProtoLengths[proto];
 
    /* Find Good Protos */
    if (Temp >= AdaptProtoThreshold) {
      *ProtoArray = proto;
      ProtoArray++;
      NumGoodProtos++;
    }
  }
 
  if (MatchDebuggingOn(Debug)) {
    tprintf("Match Complete --------------------------------------------\n");
  }
  delete tables;
 
  return NumGoodProtos;
}

Match()

void tesseract::IntegerMatcher::Match	(	INT_CLASS_STRUCT *	ClassTemplate,
		BIT_VECTOR	ProtoMask,
		BIT_VECTOR	ConfigMask,
		int16_t	NumFeatures,
		const INT_FEATURE_STRUCT *	Features,
		tesseract::UnicharRating *	Result,
		int	AdaptFeatureThreshold,
		int	Debug,
		bool	SeparateDebugWindows
	)

IntegerMatcher returns the best configuration and rating for a single class. The class matched against is determined by the uniqueness of the ClassTemplate parameter. The best rating and its associated configuration are returned.

Globals:

• local_matcher_multiplier_ Normalization factor multiplier param ClassTemplate Prototypes & tables for a class param NumFeatures Number of features in blob param Features Array of features param NormalizationFactor Fudge factor from blob normalization process param Result Class rating & configuration: (0.0 -> 1.0), 0=bad, 1=good param Debug Debugger flag: 1=debugger on

Definition at line 482 of file intmatcher.cpp.

                                                      {
  auto *tables = new ScratchEvidence();
  int Feature;
 
  if (MatchDebuggingOn(Debug)) {
    tprintf("Integer Matcher -------------------------------------------\n");
  }
 
  tables->Clear(ClassTemplate);
  Result->feature_misses = 0;
 
  for (Feature = 0; Feature < NumFeatures; Feature++) {
    int csum = UpdateTablesForFeature(ClassTemplate, ProtoMask, ConfigMask, Feature,
                                      &Features[Feature], tables, Debug);
    // Count features that were missed over all configs.
    if (csum == 0) {
      ++Result->feature_misses;
    }
  }
 
#ifndef GRAPHICS_DISABLED
  if (PrintProtoMatchesOn(Debug) || PrintMatchSummaryOn(Debug)) {
    DebugFeatureProtoError(ClassTemplate, ProtoMask, ConfigMask, *tables, NumFeatures, Debug);
  }
 
  if (DisplayProtoMatchesOn(Debug)) {
    DisplayProtoDebugInfo(ClassTemplate, ConfigMask, *tables, SeparateDebugWindows);
  }
 
  if (DisplayFeatureMatchesOn(Debug)) {
    DisplayFeatureDebugInfo(ClassTemplate, ProtoMask, ConfigMask, NumFeatures, Features,
                            AdaptFeatureThreshold, Debug, SeparateDebugWindows);
  }
#endif
 
  tables->UpdateSumOfProtoEvidences(ClassTemplate, ConfigMask);
  tables->NormalizeSums(ClassTemplate, NumFeatures);
 
  FindBestMatch(ClassTemplate, *tables, Result);
 
#ifndef GRAPHICS_DISABLED
  if (PrintMatchSummaryOn(Debug)) {
    Result->Print();
  }
 
  if (MatchDebuggingOn(Debug)) {
    tprintf("Match Complete --------------------------------------------\n");
  }
#endif
 
  delete tables;
}

Member Data Documentation

kEvidenceTableBits

const int tesseract::IntegerMatcher::kEvidenceTableBits = 9

static

Definition at line 65 of file intmatcher.h.

kIntEvidenceTruncBits

const int tesseract::IntegerMatcher::kIntEvidenceTruncBits = 14

static

Definition at line 67 of file intmatcher.h.

kIntThetaFudge

const int tesseract::IntegerMatcher::kIntThetaFudge = 128

static

Definition at line 63 of file intmatcher.h.

kSEExponentialMultiplier

const float tesseract::IntegerMatcher::kSEExponentialMultiplier = 0.0f

static

Definition at line 69 of file intmatcher.h.

kSimilarityCenter

const float tesseract::IntegerMatcher::kSimilarityCenter = 0.0075f

static

Definition at line 71 of file intmatcher.h.

---

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

• /media/home/debian/src/github/tesseract-ocr/tesseract/src/classify/intmatcher.h
• /media/home/debian/src/github/tesseract-ocr/tesseract/src/classify/intmatcher.cpp