Palmprint Recognition Based on Curvelet Transform Decision Fusion

Abstract Curvelet transform is a multi-scale method that can represent curves most sparsely. The main feature of palmprint images is that it is made up of several main curves. This paper utilizes the curvelet transform to extract the feature information of palmprint images on different scales, deals with the information by dimension reduction of PCA(Principal Component Analysis), then provides the information for RBF network to study and make decisions. Finally, according to the customers’ requirements, this paper chooses decision rules to get the fusion of the results and attain the goal of palmprint recognition. It has been proved by the experiments that this method can fully make use of the features information of images and can obtain higher recognition rate than other methods