#include <recodebeam.h>

Collaboration diagram for tesseract::RecodeBeamSearch:

[legend]

Classes
struct	RecodeBeam

Public Member Functions
	RecodeBeamSearch (const UnicharCompress &recoder, int null_char, bool simple_text, Dict *dict)

void	Decode (const NetworkIO &output, double dict_ratio, double cert_offset, double worst_dict_cert, const UNICHARSET *charset, int lstm_choice_mode=0)

void	Decode (const GENERIC_2D_ARRAY< float > &output, double dict_ratio, double cert_offset, double worst_dict_cert, const UNICHARSET *charset)

void	ExtractBestPathAsLabels (GenericVector< int > labels, GenericVector< int > xcoords) const

void	ExtractBestPathAsUnicharIds (bool debug, const UNICHARSET unicharset, GenericVector< int > unichar_ids, GenericVector< float > certs, GenericVector< float > ratings, GenericVector< int > *xcoords) const

void	ExtractBestPathAsWords (const TBOX &line_box, float scale_factor, bool debug, const UNICHARSET unicharset, PointerVector< WERD_RES > words, int lstm_choice_mode=0)

void	DebugBeams (const UNICHARSET &unicharset) const

Static Public Member Functions
static int	LengthFromBeamsIndex (int index)

static NodeContinuation	ContinuationFromBeamsIndex (int index)

static bool	IsDawgFromBeamsIndex (int index)

static int	BeamIndex (bool is_dawg, NodeContinuation cont, int length)

Public Attributes
std::vector< std::vector< std::pair< const char *, float > > >	timesteps

Static Public Attributes
static const float	kMinCertainty = -20.0f

static const int	kNumLengths = RecodedCharID::kMaxCodeLen + 1

static const int	kNumBeams = 2 * NC_COUNT * kNumLengths

Private Types
using	TopPair = KDPairInc< float, int >

Private Member Functions
void	DebugBeamPos (const UNICHARSET &unicharset, const RecodeHeap &heap) const

WERD_RES *	InitializeWord (bool leading_space, const TBOX &line_box, int word_start, int word_end, float space_certainty, const UNICHARSET *unicharset, const GenericVector< int > &xcoords, float scale_factor)

void	ComputeTopN (const float *outputs, int num_outputs, int top_n)

void	DecodeStep (const float outputs, int t, double dict_ratio, double cert_offset, double worst_dict_cert, const UNICHARSET charset, bool debug=false)

void	SaveMostCertainChoices (const float outputs, int num_outputs, const UNICHARSET charset, int xCoord)

void	ContinueContext (const RecodeNode prev, int index, const float outputs, TopNState top_n_flag, double dict_ratio, double cert_offset, double worst_dict_cert, RecodeBeam *step)

void	ContinueUnichar (int code, int unichar_id, float cert, float worst_dict_cert, float dict_ratio, bool use_dawgs, NodeContinuation cont, const RecodeNode prev, RecodeBeam step)

void	ContinueDawg (int code, int unichar_id, float cert, NodeContinuation cont, const RecodeNode prev, RecodeBeam step)

void	PushInitialDawgIfBetter (int code, int unichar_id, PermuterType permuter, bool start, bool end, float cert, NodeContinuation cont, const RecodeNode prev, RecodeBeam step)

void	PushDupOrNoDawgIfBetter (int length, bool dup, int code, int unichar_id, float cert, float worst_dict_cert, float dict_ratio, bool use_dawgs, NodeContinuation cont, const RecodeNode prev, RecodeBeam step)

void	PushHeapIfBetter (int max_size, int code, int unichar_id, PermuterType permuter, bool dawg_start, bool word_start, bool end, bool dup, float cert, const RecodeNode prev, DawgPositionVector d, RecodeHeap *heap)

void	PushHeapIfBetter (int max_size, RecodeNode node, RecodeHeap heap)

bool	UpdateHeapIfMatched (RecodeNode new_node, RecodeHeap heap)

uint64_t	ComputeCodeHash (int code, bool dup, const RecodeNode *prev) const

void	ExtractBestPaths (GenericVector< const RecodeNode > best_nodes, GenericVector< const RecodeNode > second_nodes) const

void	ExtractPath (const RecodeNode node, GenericVector< const RecodeNode > *path) const

void	DebugPath (const UNICHARSET unicharset, const GenericVector< const RecodeNode > &path) const

void	DebugUnicharPath (const UNICHARSET unicharset, const GenericVector< const RecodeNode > &path, const GenericVector< int > &unichar_ids, const GenericVector< float > &certs, const GenericVector< float > &ratings, const GenericVector< int > &xcoords) const

Static Private Member Functions
static void	ExtractPathAsUnicharIds (const GenericVector< const RecodeNode > &best_nodes, GenericVector< int > unichar_ids, GenericVector< float > certs, GenericVector< float > ratings, GenericVector< int > xcoords, std::deque< std::pair< int, int >> best_choices=nullptr)

Private Attributes
const UnicharCompress &	recoder_

PointerVector< RecodeBeam >	beam_

int	beam_size_

GenericVector< TopNState >	top_n_flags_

int	top_code_

int	second_code_

GenericHeap< TopPair >	top_heap_

Dict *	dict_

bool	space_delimited_

bool	is_simple_text_

int	null_char_

Static Private Attributes
static const int	kBeamWidths [RecodedCharID::kMaxCodeLen+1]

Member Typedef Documentation

◆ TopPair

using tesseract::RecodeBeamSearch::TopPair = KDPairInc<float, int>

private

Constructor & Destructor Documentation

◆ RecodeBeamSearch()

tesseract::RecodeBeamSearch::RecodeBeamSearch	(	const UnicharCompress &	recoder,
		int	null_char,
		bool	simple_text,
		Dict *	dict
	)

Member Function Documentation

◆ BeamIndex()

static int tesseract::RecodeBeamSearch::BeamIndex	(	bool	is_dawg,
		NodeContinuation	cont,
		int	length
	)

inlinestatic

◆ ComputeCodeHash()

uint64_t tesseract::RecodeBeamSearch::ComputeCodeHash	(	int	code,
		bool	dup,
		const RecodeNode *	prev
	)		const

private

◆ ComputeTopN()

void tesseract::RecodeBeamSearch::ComputeTopN	(	const float *	outputs,
		int	num_outputs,
		int	top_n
	)

private

◆ ContinuationFromBeamsIndex()

static NodeContinuation tesseract::RecodeBeamSearch::ContinuationFromBeamsIndex ( int index )

inlinestatic

◆ ContinueContext()

void tesseract::RecodeBeamSearch::ContinueContext	(	const RecodeNode *	prev,
		int	index,
		const float *	outputs,
		TopNState	top_n_flag,
		double	dict_ratio,
		double	cert_offset,
		double	worst_dict_cert,
		RecodeBeam *	step
	)

private

◆ ContinueDawg()

void tesseract::RecodeBeamSearch::ContinueDawg	(	int	code,
		int	unichar_id,
		float	cert,
		NodeContinuation	cont,
		const RecodeNode *	prev,
		RecodeBeam *	step
	)

private

◆ ContinueUnichar()

void tesseract::RecodeBeamSearch::ContinueUnichar	(	int	code,
		int	unichar_id,
		float	cert,
		float	worst_dict_cert,
		float	dict_ratio,
		bool	use_dawgs,
		NodeContinuation	cont,
		const RecodeNode *	prev,
		RecodeBeam *	step
	)

private

◆ DebugBeamPos()

void tesseract::RecodeBeamSearch::DebugBeamPos	(	const UNICHARSET &	unicharset,
		const RecodeHeap &	heap
	)		const

private

◆ DebugBeams()

void tesseract::RecodeBeamSearch::DebugBeams ( const UNICHARSET & unicharset ) const

◆ DebugPath()

void tesseract::RecodeBeamSearch::DebugPath	(	const UNICHARSET *	unicharset,
		const GenericVector< const RecodeNode *> &	path
	)		const

private

◆ DebugUnicharPath()

void tesseract::RecodeBeamSearch::DebugUnicharPath	(	const UNICHARSET *	unicharset,
		const GenericVector< const RecodeNode *> &	path,
		const GenericVector< int > &	unichar_ids,
		const GenericVector< float > &	certs,
		const GenericVector< float > &	ratings,
		const GenericVector< int > &	xcoords
	)		const

private

◆ Decode() [1/2]

void tesseract::RecodeBeamSearch::Decode	(	const NetworkIO &	output,
		double	dict_ratio,
		double	cert_offset,
		double	worst_dict_cert,
		const UNICHARSET *	charset,
		int	lstm_choice_mode = `0`
	)

◆ Decode() [2/2]

void tesseract::RecodeBeamSearch::Decode	(	const GENERIC_2D_ARRAY< float > &	output,
		double	dict_ratio,
		double	cert_offset,
		double	worst_dict_cert,
		const UNICHARSET *	charset
	)

◆ DecodeStep()

void tesseract::RecodeBeamSearch::DecodeStep	(	const float *	outputs,
		int	t,
		double	dict_ratio,
		double	cert_offset,
		double	worst_dict_cert,
		const UNICHARSET *	charset,
		bool	debug = `false`
	)

private

◆ ExtractBestPathAsLabels()

void tesseract::RecodeBeamSearch::ExtractBestPathAsLabels	(	GenericVector< int > *	labels,
		GenericVector< int > *	xcoords
	)		const

◆ ExtractBestPathAsUnicharIds()

void tesseract::RecodeBeamSearch::ExtractBestPathAsUnicharIds	(	bool	debug,
		const UNICHARSET *	unicharset,
		GenericVector< int > *	unichar_ids,
		GenericVector< float > *	certs,
		GenericVector< float > *	ratings,
		GenericVector< int > *	xcoords
	)		const

◆ ExtractBestPathAsWords()

void tesseract::RecodeBeamSearch::ExtractBestPathAsWords	(	const TBOX &	line_box,
		float	scale_factor,
		bool	debug,
		const UNICHARSET *	unicharset,
		PointerVector< WERD_RES > *	words,
		int	lstm_choice_mode = `0`
	)

◆ ExtractBestPaths()

void tesseract::RecodeBeamSearch::ExtractBestPaths	(	GenericVector< const RecodeNode >	best_nodes,
		GenericVector< const RecodeNode >	second_nodes
	)		const

private

◆ ExtractPath()

void tesseract::RecodeBeamSearch::ExtractPath	(	const RecodeNode *	node,
		GenericVector< const RecodeNode >	path
	)		const

private

◆ ExtractPathAsUnicharIds()

void tesseract::RecodeBeamSearch::ExtractPathAsUnicharIds	(	const GenericVector< const RecodeNode *> &	best_nodes,
		GenericVector< int > *	unichar_ids,
		GenericVector< float > *	certs,
		GenericVector< float > *	ratings,
		GenericVector< int > *	xcoords,
		std::deque< std::pair< int, int >> *	best_choices = `nullptr`
	)

staticprivate

◆ InitializeWord()

WERD_RES * tesseract::RecodeBeamSearch::InitializeWord	(	bool	leading_space,
		const TBOX &	line_box,
		int	word_start,
		int	word_end,
		float	space_certainty,
		const UNICHARSET *	unicharset,
		const GenericVector< int > &	xcoords,
		float	scale_factor
	)

private

◆ IsDawgFromBeamsIndex()

static bool tesseract::RecodeBeamSearch::IsDawgFromBeamsIndex ( int index )

inlinestatic

◆ LengthFromBeamsIndex()

static int tesseract::RecodeBeamSearch::LengthFromBeamsIndex ( int index )

inlinestatic

◆ PushDupOrNoDawgIfBetter()

void tesseract::RecodeBeamSearch::PushDupOrNoDawgIfBetter	(	int	length,
		bool	dup,
		int	code,
		int	unichar_id,
		float	cert,
		float	worst_dict_cert,
		float	dict_ratio,
		bool	use_dawgs,
		NodeContinuation	cont,
		const RecodeNode *	prev,
		RecodeBeam *	step
	)

private

◆ PushHeapIfBetter() [1/2]

void tesseract::RecodeBeamSearch::PushHeapIfBetter	(	int	max_size,
		int	code,
		int	unichar_id,
		PermuterType	permuter,
		bool	dawg_start,
		bool	word_start,
		bool	end,
		bool	dup,
		float	cert,
		const RecodeNode *	prev,
		DawgPositionVector *	d,
		RecodeHeap *	heap
	)

private

◆ PushHeapIfBetter() [2/2]

void tesseract::RecodeBeamSearch::PushHeapIfBetter	(	int	max_size,
		RecodeNode *	node,
		RecodeHeap *	heap
	)

private

◆ PushInitialDawgIfBetter()

void tesseract::RecodeBeamSearch::PushInitialDawgIfBetter	(	int	code,
		int	unichar_id,
		PermuterType	permuter,
		bool	start,
		bool	end,
		float	cert,
		NodeContinuation	cont,
		const RecodeNode *	prev,
		RecodeBeam *	step
	)

private

◆ SaveMostCertainChoices()

void tesseract::RecodeBeamSearch::SaveMostCertainChoices	(	const float *	outputs,
		int	num_outputs,
		const UNICHARSET *	charset,
		int	xCoord
	)

private

◆ UpdateHeapIfMatched()

bool tesseract::RecodeBeamSearch::UpdateHeapIfMatched	(	RecodeNode *	new_node,
		RecodeHeap *	heap
	)

private

Member Data Documentation

◆ beam_

PointerVector<RecodeBeam> tesseract::RecodeBeamSearch::beam_

private

◆ beam_size_

int tesseract::RecodeBeamSearch::beam_size_

private

◆ dict_

Dict* tesseract::RecodeBeamSearch::dict_

private

◆ is_simple_text_

bool tesseract::RecodeBeamSearch::is_simple_text_

private

◆ kBeamWidths

const int tesseract::RecodeBeamSearch::kBeamWidths

staticprivate

Initial value:

= {
    5, 10, 16, 16, 16, 16, 16, 16, 16, 16,
}

◆ kMinCertainty

const float tesseract::RecodeBeamSearch::kMinCertainty = -20.0f

static

◆ kNumBeams

const int tesseract::RecodeBeamSearch::kNumBeams = 2 * NC_COUNT * kNumLengths

static

◆ kNumLengths

const int tesseract::RecodeBeamSearch::kNumLengths = RecodedCharID::kMaxCodeLen + 1

static

◆ null_char_

int tesseract::RecodeBeamSearch::null_char_

private

◆ recoder_

const UnicharCompress& tesseract::RecodeBeamSearch::recoder_

private

◆ second_code_

int tesseract::RecodeBeamSearch::second_code_

private

◆ space_delimited_

bool tesseract::RecodeBeamSearch::space_delimited_

private

◆ timesteps

std::vector< std::vector<std::pair<const char*, float> > > tesseract::RecodeBeamSearch::timesteps

◆ top_code_

int tesseract::RecodeBeamSearch::top_code_

private

◆ top_heap_

GenericHeap<TopPair> tesseract::RecodeBeamSearch::top_heap_

private

◆ top_n_flags_

GenericVector<TopNState> tesseract::RecodeBeamSearch::top_n_flags_

private

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

/home/stephane/src/tesseract/src/lstm/recodebeam.h
/home/stephane/src/tesseract/src/lstm/recodebeam.cpp

Classes

Public Member Functions

Static Public Member Functions

Public Attributes

Static Public Attributes

Private Types

Private Member Functions

Static Private Member Functions

Private Attributes

Static Private Attributes

Member Typedef Documentation

◆ TopPair

Constructor & Destructor Documentation

◆ RecodeBeamSearch()

Member Function Documentation

◆ BeamIndex()

◆ ComputeCodeHash()

◆ ComputeTopN()

◆ ContinuationFromBeamsIndex()

◆ ContinueContext()

◆ ContinueDawg()

◆ ContinueUnichar()

◆ DebugBeamPos()

◆ DebugBeams()

◆ DebugPath()

◆ DebugUnicharPath()

◆ Decode() [1/2]

◆ Decode() [2/2]

◆ DecodeStep()

◆ ExtractBestPathAsLabels()

◆ ExtractBestPathAsUnicharIds()

◆ ExtractBestPathAsWords()

◆ ExtractBestPaths()

◆ ExtractPath()

◆ ExtractPathAsUnicharIds()

◆ InitializeWord()

◆ IsDawgFromBeamsIndex()

◆ LengthFromBeamsIndex()

◆ PushDupOrNoDawgIfBetter()

◆ PushHeapIfBetter() [1/2]

◆ PushHeapIfBetter() [2/2]

◆ PushInitialDawgIfBetter()

◆ SaveMostCertainChoices()

◆ UpdateHeapIfMatched()

Member Data Documentation

◆ beam_

◆ beam_size_

◆ dict_

◆ is_simple_text_

◆ kBeamWidths

◆ kMinCertainty

◆ kNumBeams

◆ kNumLengths

◆ null_char_

◆ recoder_

◆ second_code_

◆ space_delimited_

◆ timesteps

◆ top_code_

◆ top_heap_

◆ top_n_flags_