A Study of Novel Optical Character Recognition Algorithms

Abstract

The Optical Character Recognition has been the inseparable part of human life during everyday transaction. The OCR has extended its application areas in almost all fields viz. healthcare, finance, banking, entertainment, trading system, digital storage and so on. In the recent past, handwriting recognition is one of the hardeststudy areas in the area of image processing. In this paper the various techniques for converting textual content from number plates, printed, handwritten paper document into machine code have been discussed. The transforming method used in all these techniques is known as OCR. The English OCR system is necessary for the conversion of various published books and other documents in English into human editable computer text files. Latest researches in this area have included the methodologies thatidentify different fonts and styles of English hand written scripts. As of date, even though a number of algorithms are available, it has its own pros and cons. Since, recognition of different styles and fonts in machine printed and handwritten English script is a biggest challenge, this field is open for the researchers to implement new algorithms that would overcome the deficiencies of its predecessors. KeywordsOCR, neural network, segmentation, hidden markov, fuzzy logic