Novel Approach for Segmentation of Handwritten Touching Characters from Devanagari Words

Abstract

Segmentation in image processing refers to the process of partitioning a digital image into multiple segments. This paper makes an attempt to segment the handwritten Marathi words using Devanagari script. Segmentation of handwritten words into characters is a challenging task primarily because of complexity of structural features of the script and varied writing styles. Without dissecting these touching characters, we cannot recognize the individual characters, hence arises the need for segmentation of touching characters in a word. Paper proposes a technique which performs necessary pre-processing of a word, finds joint points in the word, identify vertical and horizontal lines and finally dissect touching characters by taking into account its dimensions namely height and width of bounding box.