Handwritten Kannada Numeral Recognition Based on Structural Features

This paper deals with the automatic recognition of handwritten Isolated Kannada numerals based on structural features. Four different types of structural features namely, directional density of pixels in four directions, water reservoirs, maximum profile distances, and fill hole density are used for the recognition of numerals. A Minkowski minimum distance criteria is used to find minimum distances and K-nearest neighbor classifier is used to classify the Kannada numerals. A total 1512 numeral images are tested, and the overall accuracy is found to be 96.12 %. The novelty of the proposed method is that it is thinning free, fast and writer style independent.