A novel classifier for handwritten numeral recognition

This paper presents a novel pattern classification approach - a kernel and Bhattacharyya distance based classifier which utilizes the distribution characteristics of the samples in each class. Bhattacharyya distance in the subspace spanned by the eigenvectors which are associated with the smaller eigenvalues in each class is adopted as the classification criterion. The smaller eigenvalues are substituted by a small value threshold in such a way that the classification error in a given database is minimized. Application of the proposed classifier to the issue of handwritten numeral recognition demonstrates that it is promising in practical applications.