Package NearOCR

Interface Summary

Interface	Description
MessageListener	An interface which allows people to receive global NearOCR events from the logger.
ProgressListener	Designed to allow objects to listen in on the progress of long OCR-related operations.
ReceptorPattern	An interface to allow creation of arbitrary receptor generating algorithms.

Class Summary

Class	Description
CharacterImage	A bitmap representation of a character.
Document	Retains information specific to one document.
DoubleObliqueReceptorPattern	Provides a double-hatching layout designed to emulate pixel layouts.
FixedLengthRandomReceptorPattern	Although slightly more capable than the random sampling receptor pattern this is still severely imperfect.
Letter	Stores a single character and forms the bulk of a Document.
Logger	The logger is a client-agnostic way of reporting events and progress from the OCR manager to the interface.
NetStack	The net stack holds all nets that are to be run durin the OCR process.
OCRManager	Manages all OCR function.
RandomReceptorPattern	A pattern of receptors with totally random position, orientation and length.
Receptor	Represents one single-unit receptor.
ReceptorManager	Manages the currently loaded receptor net and its various properties.
RegularObliqueReceptorPattern	A receptor pattern that is essentially single-hatching.
RunManager	The RunManager is responsible for all OCR running operations which actually recognise characters rather than merely adjusting weight values.
Symbol	A symbol is responsible for managing training and finding values from net output.
SymbolTable	Stores a table of all outputs correlating to their ideal output as a character and the filepath of the original image which was used to train the net, along with an average 'off' weighting.
TrainingManager	Handles all aspects relating to training techniques for both the encapsulated Neural Net and the symbol table.
WeightedNode	Holds an adjusted weight for one output.