A Probabilistic Character Computing Approach For Telugu Ocr Post-Processing

Abstract

Recent innovative technologies are highly attracted by users and reduce complicated tasks most efficiently. Conversion of digital Image documentation to editable text conversion becomes an essential tendency. OCR (Optical Character Recognition) was utilized to convert images into text. In the process of transformation, accuracy plays a vital role. In a language like Telegu, OCR fails accuracy to generate a correct word. To improve the accuracy of a word by detecting and correcting the errors, many schemes support the N-Gram model but, restricted to unigram, bigram, and trigram words. Thus, the paper proposes a novel approach called Probabilistic character computing, which computes each character of a word efficiently.