Face Recognition Method Research Base on Phase Congruency and Modular PCA
|School||Lanzhou University of Technology|
|Course||Applied Computer Technology|
|Keywords||Face Recognition Fourier transform Principal component analysis ( PCA ) Mode PCA method Phase consistency|
The face recognition technology is a very active research topic in the field of computer pattern recognition , in law, business , security systems and other fields have a wide range of applications . A rather difficult problem because of the particularity of the face image , face recognition is also the field of pattern recognition , and it is a challenging subject , it brought together a multi-disciplinary knowledge and techniques, such as signal processing , intelligent control, pattern recognition, machine vision . How to use the computer face images fast and accurate automatic identification and how to improve the recognition accuracy , image processing and pattern recognition hot and difficult , to make this technology is fully mature , there are a lot of work needs to be done . More in-depth analysis and research on existing face recognition algorithm , as well as analysis of the light of factors that affect the face recognition rate , summed up the new method based on phase consistency principle . The phase information obtained through the Fourier transform of the method contains a rich texture information , and phase consistency principle applied to the face image edge detection , and achieved very good results , and show the phase information revealed by image the essential characteristics . Experimental results show that the use of phase reconstructed image characteristics of lighting conditions can be removed to a large extent , by the light -independent features extracted face image processing to improve the recognition accuracy , but also proved that the reconstruction phase of the phase spectrum analysis of binding can be carried out light independent of the human face image feature extraction . Taking into account factors such as posture expressions to face recognition , this paper conducted in-depth research on principal component analysis ( PCA ) , and on this basis a mode PCA method based on phase congruency image . The method is both an extension of the PCA method , but also improved the traditional modular PCA method . The first face image phase consistency of treatment to eliminate the impact of light on the identification , modular but also take into account the local characteristics of the image , reducing the posture expression factors , followed by the study of face recognition algorithm provides a new way of thinking . Experimental results show that the method is effective .