EMBEDDED SPELL CHECKER FOR OPTICAL CHARACTER RECOGNITION SYSTEM FOR INDIAN LANGUAGES (HINDI

Abstract

This work is an attempt to develop "Embedded Spell Checker Module for Indian Language (Hindi)". The spell checker module developed to identify various types of errors and to correct them in Hindi text. The OCR recognizes various types of errors, which commonly occur in Devanagari. A special property of the present technique is that the multiple errors in a single word can also be corrected. The errors are detected and corrected in two steps. In first step, detection of invalid Hindi patterns is performed and in second step heuristically correction of the word followed by the dictionary look-up for generation of probable correct options is done. This Spell Checker Module checks the various errors on scanned text document. An Optical Character Recognition System takes scanned image of a text document as input and produces the document in Text format. When the OCR provides the document, document has some recognition errors. When the document is provided to this Spell Checker module, this module scans whole the document and recognizes the wrong word in the given text and corrects each and every wrong word. The software has been developed as one of the modules of 'Chitraksharika', an Optical Character Recognition system for Devanagari script being developed at ER&DCI, Noida.