AN OFFLINE CHARACTER RECOGNITION METHOD FOR PRINTED DEVANAGARI SCRIPT
Megha Nagarhalli, Aditi Nene, Aishawarya Palkar, Richa Misri, Vina M. Lomte · International journal of advance research and innovative ideas in education · 2017
Primitively, all the information stored by mankind was in manual format- textual or pictorial. Preserving, maintaining and sending manual information is a tedious job. A simple solution to this problem is converting this manual information into a digital format that also allows searching and editing. Recognition of such engraved text paves a way for making information storage independent of manpower as well as concise. Considering the popularity of Devanagari script in India, such a manual to digital conversion system has immense importance. The Devanagari script has letters with intricate details and concepts like fused letters, modifiers and shirorekha. Distinguishing between all these characteristics and making accurate analysis of text is a challenging task for engineers working in this field. The proposed system provides a simple method for tackling such issues. Letters are distinguished on the basis of continuousness of shirorekha and characters are recognised in accordance with their pixel values. These pixel values are recognised in accordance of line, word and character division. Combining results of such division, the input text can be recognised.