The Optimization Study On PCA Face Recognition Algorithm Based On Swarm Intelligence

The traditional method of identification, such as documents and password, are the largest use of identification technology, However, these identification methods are significant vulnerabilities and insecurity. The biometric face recognition technology uses a secret key biological charcteristics that can not be replicated with a non-contact, intuitive and friendly advantages. Face Recognition Based on Principal Components Analysis is one of the most classic algorithms. The PCA based on K-L, Acquiring a feature value of the image of the face image matrix by calculating the covariance matrix, selecting the larger eigenvalue corresponding eigenvectors constituting the face recognition feature subspaces, Finally, the training face image and test face image projected on the quantum space to feature face recognition than by distance. But the PCA face recognition algorithm has poor stability, low recognition rate and other shortcomings.This paper proposes an improved method for these problems. The main contents are as follows:(1) Introduces the background and significance of the study of face recognition, as well as research status PCA face recognition algorithm.(2) Introducing the image pre-processing and the basis of PCA face recognition algorithm K-L transformation and analyzes the advantages and disadvantages of PCA face recognition algorithm.(3) By introducing the Simulated Annealing and Shuffled Frog Leaping Algorithm, we improve the genetic algorithm optimization PCA face recognition algorithm and propose the genetic PCA algorithm bases on Simulated Annealing and Shuffled Frog Leaping Algorithm. The improved face recognition algorithm is verified on the ORL database. The experiments shows that the improved algorithm can avoid the algorithm "premature" and can search the global optimum value faster and improve the convergence speed and recognition rate.(4)We propose the genetic PCA algorithm which based on Binary PSO algorithm by introducing the Binary PSO. The improved face recognition algorithm is verified on the ORL database. The experiments shows that the improved algorithm can avoid the algorithm "premature" and can search the global optimum value faster and improve the convergence speed and recognition rate.The main work has been summarized in the end of this thesis, and it put forward the possible directions of the future study.