Font Size: a A A

Based Enhanced 2dpca + Pca Face Recognition System

Posted on:2011-01-06Degree:MasterType:Thesis
Country:ChinaCandidate:J H LiFull Text:PDF
GTID:2208360308971882Subject:Detection Technology and Automation
Abstract/Summary:PDF Full Text Request
Face recognition is an identification process through extracting useful information from face image by computer. Compared with other human biometric technology such as fingerprint, retina, iris, gene etc, face recognition technology has a wide range of applications on national security, civil and military security, economic and other fields because of some merits such as without excessive user participation, non-contact data acquisition, without any damage to users, easy to conceal, etc.? Now, it is becoming an important developing direction in the field of artificial intelligence and human biometric technology.Based on the construction of a static face image recognition system, this paper studies some problems about face recognition; the main work has the following four parts:(1) Face image preprocessing. Using histogram equalization, normalization method for image preprocessing in ORL face database.(2) Face image feature extracting. This paper analyzes the characteristics of principal component analysis (PCA), two-dimensional principal component analysis (2DPCA) and fisher linear discriminant analysis (FLD).According to three methods'complementary, 2DPCA+PCA enhancement method is put forward.(3) Genetic neural network classifiers designing. This paper optimizes the common BP neural network classifier by introducing genetic algorithms and designs parallel genetic neural network classifiers by improving genetic algorithms.(4) Construction of face recognition system. This paper finishes the construction of face recognition system by combining image inputting module, image preprocessing module, image feature extracting module, and neural network classification module,?also has experiments on ORL face image. Experiments show that the face recognition system based on 2DPCA + PCA enhancement has such advantages as high recognition rate and easy to program,so it has certain application value.
Keywords/Search Tags:Face recognition, Principal component analysis, Two-dimensional principal component analysis, Fisher linear discriminate analysis, Genetic algorithm
PDF Full Text Request
Related items