Biometric recognition by using audio and visual feature fusion

This paper proposes an algorithm to fuse feature information of audio and visual biometrics, and a framework of Probabilistic Neural Network (PNN) is developed to achieve feature fusion. In addition, an Audio-Visual based Text-Independent (AVTI) person recognition system has been constructed by using still visual images and text-independent audio signals. Test results show that the proposed algorithm achieves a better recognition rate, and it is superior to any single modal biometrics involved in the system. The results also indicate that the proposed method overcomes the problems of different frame rates of audio and visual signals and the curse of dimensionality normally existed in feature fusion procedure.