Zone based Feature Extraction Algorithm for Handwritten Numeral Recognition of Kannada Script

India is a multi-lingual and multi-script country, where eighteen official scripts are accepted and have over hundred regional languages. In this paper we propose Zone and projection distance metric based feature extraction system. The character /image (50times50) is further divided in to 25 equal zones (10times10 each). For each zone column average pixel distance is computed in Vertical Downward Direction (VDD) (one feature). This procedure is sequentially repeated for entire zone/grid/box columns present in the zone (ten features). Similarly this procedure is repeated for each zone from all the direction say Vertical Upward Direction (VUD), Horizontal Right Direction (HRD) and Horizontal Left Direction (HLD) to extract 10 features for each direction. Hence 40 features are extracted for each zone. This procedure is sequentially for the entire zone present in the numeral image. Finally 1000 such features are extracted for classification and recognition. There could be some zone column/row having empty foreground pixels. Hence the feature value of such particular zone column/row in the feature vector is zero. Nearest neighbor classifier is used for subsequent classification and recognition purpose. We obtained 97.8% recognition rate for Kannada numerals.