Writer identification using Gabor wavelet

This paper focuses on a system to fulfill writer identification, which identifies a person with his/her handwriting. A text independent method is proposed in this paper, which involves no local feature analysis. First, a preprocessing method is employed to normalize script images. Then, 2D Gabor wavelet technique is developed to extract the global features. Finally a K-nearest neighbor (K-NN) classifier is designed to identify the person with the identified handwriting. Illustrative experiments are made with 110 specimens of 50 people, and a result of 97.6% accuracy is achieved with the writer identification approach proposed in this paper.