Handwritten Character Recognition using Template Matching

This paper is aimed at the recognition of handwritten characters of any arbitrary size and shape. A novel template based matching scheme is employed which resulted in high matching efficiency while maintaining a low sample size for training. As the number of samples used for training is very low, the templates can be easily modified and thus this system can be simply modified to an adaptive system exploiting the uniqueness of one’s handwriting. This property turns out to be extremely useful in making better handwritten character recognition system for cell phones, PDA and tablet computers where the input is given primarily by few users. The template can be modified by including the user given input to further increase the efficiency.