Palm Vein Identification and Verification System Based on Spatial Energy Distribution of Wavelet Sub-Bands
E. Ben George · 2014
A novel personal verification and identification system using the near-infrared (NIR) images of palm veins patterns is presented in this paper. Anew approach to extract features is proposed depend on the Spatial Energy Distribution of Wavelet Sub-bands. The proposed algorithm focuses on the coefficients' values of the wavelet detail sub- bands (which, mainly reflects the edges' directions distribution in the studies region of interest). The average of the energy distribution of each sub-band is utilized to establish the feature vector. Each wavelet detail sub-band is divided into overlapped blocks, then, the coefficients' average energy is calculated for each block be used as an element in the established feature vector. The system was tested over a database collected from 250 volunteers, where 24 images are taken for the 2 palms of each person. In total, the whole database contains 6,000 images belong to 500 different palms. The attained recognition rate was excellent (100%) and the achieved minimum equal error rate (EER) is 0.068%with the high matching speed.