COMPARISON OF TWO DIFFERENT SIMILAR SPEECH AND GESTURES MULTIMODAL INTERFACES

The paper presents two different multimodal interfaces based on automatic recognition and interpretation of speech and gestures of user’s head and hands, developed within the framework of the SIMILAR European Network of Excellence. The architectures of ICANDO and MOWGLI multimodal interfaces, modalities recognition, information synchronization and fusion as well as qualitative comparison and quantitative evaluation using Fitt’s law experiments are described. The comparison of both contactless interfaces shows that in spite of the differences in computer vision and ASR techniques applied, they provide similar performances in contactless human-computer interaction. 1