Biometric Authentication using Phase Only Correlation with Compensation Algorithm for Rotation

In biometric authentication, phase only correlation is one of the effective method since the approach is robust for shift and irregularity of illumination of input images. But this method has a defect that the correlation values are considerably different when the image rotates from the original position. To improve this weak point, this paper deals with phase only correlation with compensation algorithm for rotation. For correct of rotation, we use Hough transform, Fourier transform and logarithm conversion. Our proposed method was applied to actual images of wrist vein. To capture more clearly vein pattern, we used CCD camera with infrared LEDs and enhanced it by image data processing. In this paper, those experiments in search of angle of rotation calculation precision and comparing our method with the early researches were performed. Our proposed method has possibility to be applied to identification with devices of facility, such as cellular phone