Robust real-time face recognition

Abstract

This paper describes and discusses the algorithms required to perform face detection and face recognition in real-time. Simple features, similar to Haar basis functions, are used for detection and the eigenfaces technique is used for recognition. Further to the above, a novel method of increasing face recognition rates is presented for situations where a database containing multiple images of the same subject is being used. It is shown that these well-known, existing techniques for both detection and recognition can be combined in a manner that runs in real-time, but still preserves the original success rates mentioned in literature.