Automatic Facial Expression Recognition for the Interaction of Individuals with Multiple Disabilities

Individuals affected by profound intellectual and multiple disabilities remain immobile or suffer multiple sensory and intellectual impairments. They are unable to produce conventional behaviours to communicate a particular need. Consequently, they cannot use the existing technological devices, as these tools require the understanding of conventional symbols. The INSENSION project aims to develop an intelligent platform that allows these people to interact with digital applications, with the aim of improving their quality of life. In this work, a facial expression recognition system, based on computer vision and artificial intelligence, is proposed for the INSENSION platform. The goal is to identify the facial expressions made by these people in response of an intention. The proposed system automatically recognizes diverse expressions in eyes, mouth and jaw regions, separately. Different algorithms have been trained and tested for each facial region. The experiments obtained promising results, allowing to correctly recognize the facial expressions required for persons affected by intellectual disabilities.