Research On Face Recognition Algorithm Based On Wavelet Analysis And Sparse Representation Fusion

As the most important technology in the field of biometric recognition,face recognition technology has developed for decades.With the development of technology,face recognition technology has been gradually applied to security monitoring,human-computer interaction and distance education and other fields.Face recognition technology that has application value can effectively extract face feature from high-dimensional face data,and suppress the interference of illumination,facial expression and occlusion on recognition performance.At the same time,it also requires precise recognition performance and efficient recognition efficiency.It can be said that face recognition technology is a long-term and difficult task.When the face image is disturbed by illumination shadow,the illumination invariant feature obtained by the illumination invariant extraction algorithm based on wavelet transform will have the effect of pseudo contour.Therefore,based on the original algorithm,this paper combines the approximate axisymmetric face preprocessing method to eliminate the false contour phenomenon of the face,and finds the suitable combination form of wavelet base and shrinkage parameters in different face database by using experimental comparison.The recognition performance of the algorithm is optimal.Finally,the comparison experiments with other algorithms show that the recognition performance of this algorithm is better than that of other algorithms.The weighted nearest neighbor sparse representation algorithm can solve the problem that the traditional sparse representation algorithm is weak for the classification of nonlinear samples,but the sparse degree of the reconstruction coefficients of the test samples obtained from the weighted matrix is not enough.Therefore,by transforming negative exponential function,this paper gives an invalid weight value to the training sample which is far away from the test sample,and a larger weight value to the training sample that is close to the test sample.The reconstruction coefficients obtained by the improved algorithm become more sparse.Finally,the comparison experiment shows that the improved algorithm has better classification and recognition performance.A face recognition algorithm based on wavelet fusion and sparse representation is proposed.In order to obtain more abundant face feature information,Gao Si smoothing filter is used to extract the feature of low-frequency component,and fusionwith the illumination invariant feature of high-frequency component as the final feature information.Then using the improved sparse face classifier for classification recognition.Through the simulation experiments on different face databases,the most suitable wavelet bases are selected,and it is proved that the sparse classifier has the best classification and recognition effect.Finally,the experimental results show that this algorithm has better classification and recognition performance.