NearOCR

Class OCRManager

java.lang.Object

NearOCR.OCRManager

public class OCRManager
extends java.lang.Object

Manages all OCR function. This is done in an interface-agnostic way.

Version:

0.1

Author:

Stephen Wattam

Constructor Summary

Constructors

Constructor and Description
OCRManager() Creates a new OCR manager, with default original values.

Method Summary

Methods

Modifier and Type	Method and Description
void	addSymbol(Symbol p_symbol) Adds a symbol to the symbol table.
void	analyse(double covarianceWeight, double subtractionWeight, ProgressListener listener) Analyses all letters by running them through all nets.
void	analyseSymbolWeights(ProgressListener listener) Analyses and updates all symbol weights to compensate for individual net idiosynchrasies.
void	appendNet(NeuralNet net) Appends a net to the stack.
int	countInputs() Returns the number of inputs as shared across the net stack.
int	countOutput() Returns the number of outputs in the final layer of the entire net stack.
Document	getDocument() Returns the document that this OCRManager is using.
NeuralNet	getNet(int p_index) Returns a net from the given index number.
NeuralNet[]	getNets() Returns all nets from the net stack, is merely a wrapper for the Net stack object.
ReceptorManager	getReceptorManager() Returns the current ReceptorManager that will be used for analysis.
SymbolTable	getSymbolTable() Returns the entire current sumbol table.
void	loadImage(java.io.File p_file) Sets the current document's image source.
void	loadSymbols(java.io.File f) Loads a symbol table from the given file.
void	moveNet(int p_indexFrom, int p_indexTo) Moves a net in the stack.
void	removeNet(int p_index) Removes a net from the internal list.
void	removeNet(NeuralNet net) Removes a net from the stack.
void	setAllWeights(double p_weight) Resets all symbol weights to the given number.
void	setDocument(Document p_doc) Sets the document to parse.
void	setReceptorManager(ReceptorManager p_manager) Sets the receptor manager.
void	slice(double p_thresholdX, double p_thresholdY) Attempts to find letters by dividing an image up by lines and letters.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

OCRManager

public OCRManager()

Creates a new OCR manager, with default original values.

Method Detail

countOutput

public int countOutput()

Returns the number of outputs in the final layer of the entire net stack.

Returns:

The number of outputs in the final layer of the net stack

setReceptorManager

public void setReceptorManager(ReceptorManager p_manager)

Sets the receptor manager.

Parameters:

p_manager - The ReceptorManager to use for analysis.

getReceptorManager

public ReceptorManager getReceptorManager()

Returns the current ReceptorManager that will be used for analysis.

Returns:

The current ReceptorManager

removeNet

public void removeNet(int p_index)
               throws java.lang.IndexOutOfBoundsException,
                      ItemNotFoundException,
                      ValueOutOfBoundsException

Removes a net from the internal list.

Parameters:

p_index - The index number of the net to remove

Throws:

IndexOutOfBoundsException - In the event that no net exists

java.lang.IndexOutOfBoundsException

ItemNotFoundException

ValueOutOfBoundsException

getNets

public NeuralNet[] getNets()

Returns all nets from the net stack, is merely a wrapper for the Net stack object.

Returns:

All nets from the stack, in order, as an array

getNet

public NeuralNet getNet(int p_index)
                 throws java.lang.IndexOutOfBoundsException

Returns a net from the given index number.

Returns:

The net at the index given in the net stack

Throws:

java.lang.IndexOutOfBoundsException

countInputs

public int countInputs()

Returns the number of inputs as shared across the net stack.

Returns:

The number of inputs

moveNet

public void moveNet(int p_indexFrom,
           int p_indexTo)

Moves a net in the stack.

Parameters:

p_indexFrom - The index of the net to move

p_indexTo - The index where the net will be placed

removeNet

public void removeNet(NeuralNet net)
               throws ItemNotFoundException,
                      ValueOutOfBoundsException

Removes a net from the stack.

Parameters:

net - The neural net to remove from the net stack.

Throws:

ItemNotFoundException

ValueOutOfBoundsException

appendNet

public void appendNet(NeuralNet net)
               throws ValueOutOfBoundsException

Appends a net to the stack.

Parameters:

net - The net to append to the net stack.

Throws:

ValueOutOfBoundsException - in the event that the number of inputs does not match that of the other loaded nets.

addSymbol

public void addSymbol(Symbol p_symbol)

Adds a symbol to the symbol table. The order of the symbol table is technically an irrelevance, as each symbol retains its own weighted edges which are automatically adjusted to fit the current net.

Parameters:

p_symbol - The symbol to add to the currently loaded symbol table

loadSymbols

public void loadSymbols(java.io.File f)
                 throws java.io.IOException

Loads a symbol table from the given file.

Parameters:

f - The file to load a symbol table from.

Throws:

java.io.IOException - In the event that any error occurs during loading

getSymbolTable

public SymbolTable getSymbolTable()

Returns the entire current sumbol table.

Returns:

The currently loaded symbol table.

getDocument

public Document getDocument()

Returns the document that this OCRManager is using.

Returns:

The document that this OCR manager is based on.

setDocument

public void setDocument(Document p_doc)

Sets the document to parse.

Parameters:

p_doc - The document to set as the current document in this OCR manager.

loadImage

public void loadImage(java.io.File p_file)
               throws java.io.IOException

Sets the current document's image source.

Parameters:

p_file - The image file to read into the document.

Throws:

java.io.IOException

analyse

public void analyse(double covarianceWeight,
           double subtractionWeight,
           ProgressListener listener)

Analyses all letters by running them through all nets. The return type is void because all output gets written back into the individual Letters' value fields.

Parameters:

listener - A ProgressListener that is to watch the progress of the operation, which could take some time.

setAllWeights

public void setAllWeights(double p_weight)

Resets all symbol weights to the given number.

Parameters:

p_weight - The number to set all weights to

analyseSymbolWeights

public void analyseSymbolWeights(ProgressListener listener)

Analyses and updates all symbol weights to compensate for individual net idiosynchrasies.

Parameters:

listener - A ProgressListener that is to watch the progress of the operation, which could take some time.

slice

public void slice(double p_thresholdX,
         double p_thresholdY)

Attempts to find letters by dividing an image up by lines and letters. This populates letters in the current document, which can be retrieved using other methods or straight from the document.

Parameters:

p_thresholdX - The X-direction threshold for cutting up images (switches the tracer on/off based on averages values

p_thresholdY - The Y-direction threshold for cutting images up