Regresar

Advanced Metrics to Evaluate Autistic Children’s Attention and Emotions from Facial Characteristics Using a Human-Robot-Game Interface

Abstract:

Several studies have discussed how to support children with autism spectrum disorder (ASD) using technological resources and the difficulties in their social interaction and non-verbal communication. In recent years, some authors have explained that a way to treat ASD syndrome has been through robot-assisted therapy, expecting to capture their attention. This study analyses aspects of the results obtained using the social robot ‘LOLY,’ which allows the children’s interaction with games, recorded for therapeutic and behavioral analysis. Aiming to refine the metrics used to automatically assess the degree of attention and emotions of children with ASD when interacting with a social robot, several human-robot interaction sessions were analyzed in the laboratory and the field. The results of the automatic processing of the video signals corresponding to the interactions were compared with classical observational techniques. Part of the analysis will also consist of comparing the results of the videos taken from different reference points of the robot, such as from the head and bust. As a result of the trial-error follow-up, the comparative analysis between automatic and observational techniques proves the effectiveness of these adjustments.