Fingerprint classification by Block Ridgelet and SVM
Amina Serir, Farida Bennabes · 2010
The present article focuses on the classification of fingerprints. Our aim goal is to unify the process of fingerprint compression, classification and identification. The well known methods suited to these tasks are based on WSQ (Wavelet Scalar Quantization) for compression, Gabor filters for classification and minutiae matching for identification. We propose to use Block Ridgelet Transform (BRT) to characterize the local structures for compression, classification and identification. This paper turns on the fingerprint classification by combining characteristics from the BRT to a Support Vector Machine (SVM) classifier. Our method allows both to easily increase the number of orientation in the analysis and the size of the descriptor. The design and implementation of ridgelet classification scheme are discussed. In order to evaluate the performance of the algorithm, FCV2002, FCV2004 fingerprint databases have been considered. The results show that this method has a serious potential in fingerprint classification.