Study on Checking Attendance System Based on Face Recognition

In this paper, the core algorithm is studied, which aiming at developing a robust and practical checking attendance system based on face recognition. In order to solve recognition rate reduction resulting from illumination, pose etc variations to some extent, the Kernel Fisher Discriminant analysis (KFD) method is studied, which the corrected model and the fractional power polynomial kernel function are employed to further improve robustness to the above variations. Through the Matlab simulation experiments on ORL face database, it is showed that the method is of higher recognition rate and more robustness to the above variations.