The Study On Face Recognition Model Based On Sparse Representation

With the rapid development of information society,the information security is more important than ever.Traditional security methods,such as password,personal identification code and key cann't fulfill the requirement of information security.The identity identification based on biometrics using the inherent physiological characteristics or behavioral characteristics of the human body has unique advantages over other methods.Biological characteristics are the unique feature of human beings.The use of biometrics for identification can fundamentally eliminate forgery and theft,with higher reliability,safety and usability.It can satisfy the requirement of national public safety,social security and financial security.Among the recognition technology based on biometrics,face recognition using the most unique physiological characteristics of humans has been attached much attention.Unfortunately,traditional face recognition technology is sensitive to illumination,expression,posture and so on.In response to these challenges,scholars have make great effort in recent decades and make fruitful achievements.Recently,the face recognition based on sparse representation has realized the face classification and recognition by using the sparse coding of the face in the training dictionary,which has been widely researched by the researchers for its robustness.This paper studies the face recognition based on sparse representation.Firstly,the sparse framework of face recognition is introduced.In the framework,the feature extraction method is used to acquire robust face feature under non-ideal conditions.Then the feature data is mapped into low dimensions in order to effectively reduce the amount of data and the computational overhead,and the sparse representation classification is used for face recognition with performance guarantees.Under the principle of sparse face recognition,the sparse face recognition system is designed and developed.The system is divided into several modules:pre processing module,data reduction module and face sparse representation module.According to the related theories,the three modules are implemented and optimized.The experiment in the ORL and Yale face database verified the effectiveness and superiority of the sparse face recognition.Sparse representation for face recognition is the frontier developing theory,relevant theories and methods need to be further improved.The Study of this paper provided a workable technology for face recognition under non-ideal conditions,promoting the development of the sparse face recognition technology in engineering applications.