Person Identification System in a Platform for Enabling Interaction with Individuals Affected by Profound and Multiple Learning Disabilities

Individuals affected by profound and multiple learning disabilities have severe restricted mobility and are subject to multiple sensory and intellectual impairments. They are therefore unable to produce conventional behaviors that may serve to communicate a particular need to a caregiver. Within the INSENSION project, an intelligent platform for enabling the interaction of this kind of people with others, is planned to be developed. The final goal is to increase their ability of self-communication through digital services and enhance the quality of their life. The system will recognize facial expressions, body gestures, vocalizations and physiological parameters using the information captured by a set of video-cameras and sensors. Based on previous knowledge about each person, the platform will be able to associate the recognized expressions with their meaning in an individualized way. For this reason, a first stage of person identification is highly required in order to personalize the understanding. In this work, a new facial recognition method is developed and properly configured to be included in the INSENSION platform. The proposed system is able to identify six individuals as well as discard the other people that could appear in the videos, assuring the monitoring of the right person.