k-nearest neighbor based offline handwritten Gurmukhi character recognition

Abstract

Offline handwritten character recognition has been a frontier area of research for the last few decades under pattern recognition. Recognition of handwritten characters is a difficult task owing to various writing styles of individuals. A scheme for offline handwritten Gurmukhi character recognition based on k-NN classifier is presented in this paper. The system first prepares a skeleton of the character, so that feature information about the character is extracted. There is abundant literature on the handwriting recognition on non-Indian scripts, but there are very few article available related to recognition of Indian scripts such as Gurmukhi. This paper presents an efficient offline handwritten Gurmukhi character recognition system based on diagonal features and transitions features using k-NN classifier. Diagonal and transitions features of a character have been computed based on distribution of points on the bitmap image of character. In k-NN method, the Euclidean distance between testing point and reference points is calculated in order to find the k-nearest neighbors. In this work, we have taken the samples of offline handwritten Gurmukhi characters from one hundred different writers. The partition strategy for selecting the training and testing patterns has also been experimented in this work. We have used in all 3500 images of Gurmukhi characters for the purpose of training and testing. The proposed system achieves a maximum recognition accuracy of 94.12% using diagonal features and k-NN classifier.