期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2015
卷号：74
期号：2
出版社：Journal of Theoretical and Applied
摘要：Biological features such as face, fingerprint, palm print and iris are widely used for the identification of a human. Due to low cost acquisition devices and high accuracy, fingerprint recognition is broadly used. Fingerprint is unique and pattern unchanged during the life time. A novel approach for fingerprint recognition system based on Gabor Wavelets and Dynamic time Warping is proposed in this work. Firstly, image enhancement has done in both spatial and frequency domains with Histogram Equalization and Fast Fourier Transform. 24 optimized Gabor kernels are generated in spatial and frequency domains and those are invariant to 6 orientations and 4 scales. In this work, the performance of the proposed method is evaluated on VFR and FVC-2006 databases. Based on the constraints of Dynamic Time Warping, an optimal warp has identified among the feature vectors and achieved 94.2% and 83.6% recognition rates on VFR and FVC-2006 databases respectively.
关键词：Fingerprint Recognition; Histogram Equalization; Fast Fourier Transform; Gabor Wavelets; Dynamic Time Warping
