Farzam Kharajinezhadian - Faculty of Biomedical Engineering, Islamic Azad University, Science and Research branch
Saeid Rashidi - Faculty of Biomedical Engineering, Islamic Azad University, Science and Research branch

Various features such as principal lines, wrinkles and edges with high acceptability led to palmprint recognition has drawn attention from researchers. Research on feature extraction can be classified into three categories: 1)Line-based, 2)subspace-based, 3)texture-based. In this paper, we consider the palmprint as a texture and apply 2D-Gabor filters and discrete wavelet transform for feature extraction. Features are classified with new approach and using K-Nearest Neighbor, Support Vector Machine, Parzen Window and Fuzzy K-Nearest Neighbor classifiers. In CASIA testing database of 5,502 palmprint samples from 312 palms, we achieved Equal Error Rate of 6.45% 0.37 and Accuracy of 93.55% 0.37 with K-Nearest Neighbor classifier

biometric, palmprint, Gabor filter, wavelet