No abstract

A feature extraction algorithm of near-infrared human face based on wavelet transform and 2DPCA (two dimension principal component analysis) was proposed. This algorithm firstly applied wavelet transform to the near infrared face images, obtained the low frequency components of face images, and removed high frequency components. Secondly applied 2DPCA to the low frequency components of the face images for feature extraction. Finally completed the face recognition using Euclidean distance .The experimental results based on near-infrared face database clearly showed that the proposed algorithm could get higher recognition rate than the traditional PCA and 2DPCA algorithm, which demonstrated the efficiency of the proposed method.

Wavefront coding technology can extend the depth of field of the iris imaging system, but the iris image obtained through the system is coded and blurred and can't be used for the recognition algorithm directly. The paper presents a fuzzy iris image restoration method used in the wavefront coding system. After the restoration, the images can be used for the following processing. Firstly, the wavefront coded imaging system is simulated and the optical parameter is analyzed, through the simulation we can get the system's point spread function (PSF). Secondly, using the blurred iris image and PSF to do a blind restoration to estimate a appropriate PSFe. Finally, based on the return value PSFe of PSF, applying the regularization filter on the blurred image. Experimental results show that the proposed method is simple and has fast processing speed. Compared with the traditional restoration algorithms of Wiener filtering and Lucy-Richardson filtering, the recovery image that got through the regularization filtering is the most similar with the original iris image.