Research On Multimodal Biometric Authentication Using Face And Palmprint

Biometric authentication technology is an identity authentication technology which makes use of physiology characteristics like face, iris, fingerprint and behavior characteristics like gait to identify individuals. In recent years, owning to the emphasizing to security, Biometric authentication technology obtained full-grown development. People pay more and more attention to it and it going deeply into our daily life. Domains, like public security, customs, financial are all applied with biometric authentication technology. But single modal authentication technology has problems like non-universality, invalidity and inaccuracy, multimodal biometric authentication technology gradually becomes one of biometric authentication research focuses. Multimodal biometric authentication technology is a new technology which uses many kinds of biometric characteristics to identify individual. Because of combination diversity and fusion tactics richness, multimodal biometric authentication technology overcomes shortages of single biometric authentication in essence then it implements more robust identity authentication system. Currently, the effectiveness of multimodal biometric authentication technology received attention and acceptance.This work investigates multimodal biometric authentication system and makes use of face and palmprint biometric characteristics to fuse these biometric modals in feature level. I use different feature extracting methods for face and palmprint modals, then fuse them by two different weight methods. By doing experimental comparison to two different weight fusing methods, I intend to find a multimodal biometric authentication system which can get higher recognition rate and more stable performance.The main content of this work is face and palmprint feature extracting methods and different weight methods. The definite contents are as follows, First, this work discusses face and palmprint feature extracting methods and adopts different methods for them.Second, the effectiveness of fuse is the key for recognition. This work respectively uses traversal weighting method and user-specific weighting method to fuse face and palmprint features.Experiments show that multimodal biometric authentication I adopted is better than single modal authentication. The user specific weighting method is an effective multimodal fusion method.