Towards a Methodology to Support Augmentative and Alternative Communication by means of Personalized Gestural Interaction.

Augmentative and Alternative Communication (AAC) involves the use of non-verbal modes as a complement or substitute for spoken language, supporting communicative abilities of people, especially people with speech limitations. Computing systems have been proposed to support AAC, applying different technology to address usersu0027 different needs. Computer vision techniques can assist people with motor impairments by using their remaining functional motions. This paper proposes a methodology to support AAC of people with motor impairments, using computer vision and machine learning techniques to enable personalized gestural interaction. The methodology was instantiated in a pilot system described in this paper and evaluated by Human-Computer Interaction experts. The evaluation results suggested improvements for the methodology and for the system, and indicated the methodology is feasible to support the design of AAC systems, and that the developed system is promising to support AAC.