A hybrid face recognition system based on multiple facial features

Abstract

In the last decade, face recognition has played an important role in many applications such as public security systems, criminal identification, video surveillance, and the identity verification for network transaction platforms. In this paper, we propose a new hybrid scheme to improve the face recognition rate. The Adaboost algorithm is employed to detect the whole face and some specific facial areas such as eyes, nose, and mouth in a grayscale image. By using the principal component analysis, the proposed system efficiently integrates the features obtained from the whole-face and from the specific facial areas. Each feature record is created and stored in a feature space to form a feature database. In the identification procedure, test images are projected onto some feature spaces and then compared with the ones stored in the database. To reduce the influence of facial expression, the proposed scheme adjusts the weights according to the importance of the whole face and facial areas. Finally, it applies a voting procedure with the support vector machine and the Euclidean distance to obtain the optimized recognition results. According to the experimental results, the proposed system provides a recognition rate of as high as 99.16%, which significantly outperforms previous methods.