Handwritten Gurmukhi Character Recognition

This paper deals with the offline recognition of handwritten Gurmukhi characters. Here two sets of features based on gradient and curvature of character image are computed. The extracted features are then fused together to form a composite feature vector containing both gradient and curvature information. Two ways of generating this composite feature vector is presented in this paper. Dimensionality of the generated composite feature vectors is set to 400. The efficiency of these feature sets is tested on a handwritten database of Gurmukhi characters containing 7000 sample character images. The experimental result demonstrates the usefulness of curvature-based feature guided by gradient information and recognition rate of 98.56% is obtained. Support Vector Machine (SVM) is used for classification purpose.