HHBCR: A novel approach for Hand-written Hindi Barakhadi Characters Recognition

Abstract

 The process that enables identification and classification of text characters from the image is termed as character recognition. This recognition is further categorized into two categories: (1) recognition of printed characters, and (2) recognition of hand-written characters. Hand-written Hindi character recognition is the process of recognizing Hindi characters from the image containing hand-written text. The work discussed in literature, related to hand-written Hindi character recognition, focuses on recognition of an individual vowel character or consonant character. As per our observations, this existing work possesses two limitations: (1) It does not focus on the recognition of characters formed by the combination of vowels and consonants known as Barakhadi characters, and (2) It recognizes individual characters less accurately because of similarity in the shapes of various characters. Hence, with an aim to overcome these limitations, in this paper, we propose a novel approach HHBCR for recognition of hand-written Hindi Barakhadi characters. Experimental results affirm recognition of characters with an acceptable accuracy.