A brief review on integrated audio-visual processing for personal identification

Personal identification has potential applications in numerous areas where the identity of a person needs to be determined or validated. The use of human identification in personal data systems used by organisations is remarkably little-discussed in the information systems literature. This paper presents a brief review on speaker identification, face recognition and integrated audio-visual systems for personal identification. A brief literature review on existing systems based on voice, face and integrated audio-visual cues is given, together with remarks on the advantages of using integrated audio-visual systems. It appears that a lot of work has been done on speaker identification and face recognition but not much on combining the two approaches in one system. In general, the voice and face of a person are sources of information presumed to uniquely identify a person. It is well accepted that the likelihood for the voices or faces of two persons to be identical is extremely small. The integration of audio and visual data for personal identification has the advantages of higher recognition performance than speaker recognition or face recognition systems operating separately. Recognition performance is expected to improve in terms of accuracy and robustness against variation of either facial or vocal information