Handwritten Kannada character recognition based on Kohonen Neural Network

In general, Online Handwriting Recognition refers to the dynamic movement of a Digitized pen on touchpad which simply involves collection of a sequence of x-yco-ordinates used to describe the online handwriting data. This paper presents a novel approach for online handwriting recognition of Kannada characters by combining Direction based Stroke Density principle(DSD) with Kohonen Neural Network (KNN). DSD principle forms the basis for feature selection whereas the subsequent Classification stage is carried out by KNN. The proposed method has been tested for 49 characters and 10 numerals of Kannada Language with 20 different handwritings resulting in an accuracy of 94.4%. This method is simple to implement and realize, also it is computationally efficient.