Algorithm of Face Authentication Based on Wavelet Transform and Combining Two Different 2-Dimensional Participial Component Analyses
Yonghui Wang · Journal of Shenyang Jianzhu University · 2010
Identity verification based on faces,as a kind of biometric authentications,has become one of the hot spots on research in the field of identity verification.In this paper,an algorithm of face verification which is based on wavelet transform and combining two different 2-Dimensional Principle Component Analyses(2DPCAs)is proposed.At first,face image is decomposed by 2-Dimational Wavelet Transform(2DWT).Then two different 2DPCAs are used to get the results of face recognition respectively.To the end,the final result is given by fusing them.Experiment on ORL database shows that the algorithm is valid.According to the results,the best recognition rates are 99.92% and 94.25% in the case of small-sample and all-sample,respectively.Algorithm of face verification based on wavelet transform and combining two different 2DPCAs is superior to that of 2DPCA,single wavelet transformation 2DPCA and 2DICA.It also maintains certain stability within changes of expression,pose and illumination.