Fusion of multiple experts in multimodal biometric personal identity verification systems

Author

Kittler, J. ; Messer, K.

Author_Institution

Centre for Vision, Speech & Signal Process., Surrey Univ., Guildford, UK

fYear

2002

fDate

2002

Firstpage

3

Lastpage

12

Abstract

We investigate two trainable methods of classifier fusion in the context of multimodal personal identity verification involving eight experts which exploit voice characteristics and frontal face biometrics. As baseline classifier combination methods, simple fusion rules (Sum and Vote) which do not require any training are used. The results of experiments on the XM2VTS database show that all four combination methods investigated yield improved performance. Trainable fusion strategies do not appear to offer better performance than simple rules.

Keywords

biometrics (access control); expert systems; face recognition; image classification; knowledge based systems; speech recognition; Sum fusion rule; Vote fusion rule; XM2VTS database; baseline classifier combination methods; behavior knowledge space; classifier fusion; decision templates; frontal face biometrics; fusion rules; multimodal biometric personal identity verification; multiple experts fusion; trainable fusion strategies; trainable methods; voice characteristics; Biomedical signal processing; Biometrics; Cameras; Data security; Face detection; Fingers; Iris; Speech processing; Surveillance; Voting;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks for Signal Processing, 2002. Proceedings of the 2002 12th IEEE Workshop on

Print_ISBN

0-7803-7616-1

Type

conf

DOI

10.1109/NNSP.2002.1030012

Filename

1030012