Feature extraction and recognition of face image based on 2DPCA with LDA algorithm

Face recognition has become a hot research topic in many fields such as biometric recognition and pattern recognition. By processing face images, different features of faces are extracted for matching and recognition. To solve the "Small Sample Size'' problem caused by LDA algorithm, this paper combined 2DPCA with LDA algorithm and its improvement algorithm 2DLDA to achieve feature extraction and recognition of human faces. In order to improve the disadvantage of slow recognition time of 2DPCA with LDA algorithm, 2DPCA combined with PCA and LDA improvement algorithm was proposed. The experimental results illustrated that the improved algorithms, 2DPCA combined with 2DLDA and 2DPCA combined with PCA and LDA, which perfectly avoided the "Small Sample Size'' problem, can extract the face features faster and improve the recognition efficiency effectively.