Face Recognition Algorithm Of Sparse Representation Based On Dictionary Transform

With the development of intelligence cities and smart life,face recognition has gradually been entering every area of people's lives.Face recognition which is applied on traffic security,college management,and banking systems makes life not only convenient and fun,but also interactive intelligence.The facial recognition algorithm based on sparse representation is attracted much attention.On the relevant domestic and foreign scholars study basis,three algorithms have been proposed to improve the face recognition algorithm which is based on sparse representation.Firstly,in view of the sparse representation-based classification(SRC)focusing on finding the best linear combination of the training samples from all classes that is approximately equal to the test sample,the algorithm of coarse to fine coefficient accumulation face recognition based on sparse rules is proposed.Firstly using sparse rules the sparse coefficient accumulation of the most two adjacent categories is acquired.Then according to the relative difference of sparse coefficient accumulation,we perform the soft constraints for all training samples and obtain the candidate classes,complete coarse to fine selection process and select the different way of classification.Finally the results of two kinds of classification model are combined and the sparse coefficient accumulation of each class is calculated for the final classification.Secondly,considering of the ordinary training samples dictionary learning unused the common information of all classes,the common space and class-wise remain space which is related categories is introduced,and hybrid sparse representation is proposed for common basis and the basis of class-wise remain based face recognition algorithm.Firstly the unlabeled common basis is learned from all kind of training samples with PCA and its reconstruction samples with common information also are acquired.Then taking advancing of origin training samples and reconstruction samples,the basis of class-wise remain with discriminative information is constructed.Finally the hybrid dictionary is obtained by combining common basis and the basis of class-wise remain and then the test sample is classified by using of residual error SRC criterion.Finally,the set-to-set distance based methods ignores the relationship between gallery sets,while representing the query set images individually over the gallery sets ignores the correlation between query sets images.In view of multiple representations of images contributing to providing complementary information,hull of image set based collaborative representation for face recognition is proposed.Firstly the extended image set with multiple representations can be obtained by jointing the domain images of the original images and mirror images and pixels with moderate intensities images.Secondly the extended dictionary is modeled as hull dictionary with non-parametric approaches for image set modeling and the query set from the same class of different domain image sets is modeled as a hull.The idea of collaborative representation and iterations are used to solve the coefficient of hull.Finally the query set is classified by using of residual error SRC criterion.