Research Of Several Effective Feature Extraction Algorithms

The technology of human face recognition is an active subject in the field of pattern recognition. There are broad application in the fields of laws, business and safety systems. For the particularity of human face images,face recognition with a computer is a very difficult problem and there are still many works to do before such technology can be used wildly. With the development of the society,the application of face recognition systems will be wilder and brings much challenge to the researchers.Face recognition system is a kind of pattern recognition system based on information processing. It can be divided into two parts: feature extraction and pattern classification. The first part is to find out a set of features that can represent the images from different persons; the second part classifies the features got from the first part. The performance of the system depends on both of the parts.For the part of feature extraction,we use the methods based on second-order and high-order square of statistics recognition. In this paper we have introduced classical face recognition methods in present,including Principle Component Analysis and Fisher face. In addition, we have introduced the pattern recognition methods based on kernel tricks and the methods directional based on images: two-dimentional PCA and two-dimentional LDA.The old methods didn't take the quality of training samples in to account. if training samples have errors,they will mislead the training of classifier, forming bad projection direction. Being enlightened by Fisher theory,in this paper we have put forward a method to adjust training samples to reduce the influence of training errors.Face recognition is a pattern recognition using little training samples. how to use the training samples we have to create more training samples to extract enough feature is a problem we should solve. In this paper we put forward a new method to create more training samples. Experimentation have proved the method is efficiency.When we use Fisher face, the scatter matrix must be singularity. In this...