Research Of Face Recognition Method Based On LDP

Human face is a kind of evidence that human beings are born and can be used as information exchange of human in the network age.It is independent,unique and not replicable,but also has various three-dimensional information,including visual and multi-dimensional Identification information.Therefore,facial recognition technology is gradually used in the Internet age,and iris recognition,fingerprint recognition and other biological unique features compared to its unique advantages,mainly for the high concealment,good concurrency,non-contact,right Hardware requirements are low.On the basis of summarizing the achievements of predecessors,this paper deeply studies the face recognition method based on local direction mode(LDP).The main research work includes:(1)According to the existing methods,only the LDP features of the image are used,and the face recognition method based on LDP and Bayesian model is proposed without using the shortcomings of human a priori information.Firstly,we study the prior information of the similarity of the LDP histogram of the similar sample image and the heterogeneous sample image in the independent training set,and estimate the conditional probability density function(Similar to the heterogeneous sample respectively).Secondly,the face image The LDP histogram compares whether the image is the probability value of a certain type of image;the last is classified using the Bayesian rule.This method has been used to identify experiments on ORL and Yale face database.Compared with traditional PCA,LBP and LDP methods,the face recognition rate has been significantly improved.(2)In order to further improve the recognition rate of traditional LDP method,an integrated classifier face recognition method based on DCT and LDP feature is proposed in combination with the whole feature,local feature and the advantages of integrated classifier.This method takes into account the advantages of DCT and LDP,and uses the idea of decision fusion to realize the integrated face recognition method.The simulation results show that the result of ORL and Yale face database has a significant improvement in the face recognition rate compared with the traditional PCA,LBP and LDP algorithms.