Multimodal user identification in a intelligent robot environment

The attempt to recognize a user through the single recognition expert like the face recognition and speaker recognition is actively comprised. But these single recognition techniques are difficult to cover the user recognition perfectly since the real environment in which a robot is positioned include the variations of the illumination and noise, and etc. In this paper, the fusion technique using face and speaker information is proposed. The proposed approach is concerned with the aggregation of the individual classifiers by means of the fuzzy integral. This stage includes two steps such as (a) the determination of the fuzzy sets described the classification results of each classifier, and (b) the fuzzy measure itself which reflects the relevance of the classifiers. The experimental results obtained for the ETRI-HRI databases reveal that the approach presented in this study yields better classification performance in comparison to the results produced by other classifiers and fusion methods.