Towards the Automatic Construction of Multimodal Graphical and Voice Interfaces

The development of natural language processing is popularizing the use of voice interfaces in current applications. However, many of the interfaces currently developed are graphical interfaces that have been surpassed for the new needs of users. Hence, the use of multimodal interfaces, both visual and voice, is necessary to be more friendly and easy to use by final users. To do this, it would be necessary to redesign current graphic systems to integrate voice interfaces. This work shows that it is possible to develop multimodal interfaces starting from existing graphical interfaces by adding voice interfaces through previously designed dialogue and interactions. Tests were carried out through Web forms using feature extraction techniques, where satisfactory results were obtained.