A novel clustering based fuzzy approach for character segmentation in handwritten odia scripts

Character Segmentation is a pivotal decisive component in Natural Language Processing that seeks to dissolve an image of an array of characters into sub-images of distinct symbols. Literature survey reveals that very little work has been reported in character segmentation of handwritten Odia documents. In this paper, a novel clustering based Fuzzy schema is presented for segmentation of characters in unconstrained Odia language. The simulation based experiments of proposed methodology are conducted on transcribed Odia scripts. It is observed that the segmentation performance in terms of Success Rate and F-Measure parameters is superior to the other existing state of the art segmentation techniques.