Intelligent Emotion Prediction System for Help in Telemedicine Therapies of Children with ASD

This paper proposes the creation of an intelligent system capable of differentiating between basic emotions as a result of identifying micro expressions in children with Autism Spectrum Disorder (ASD). This system will be developed with the aim of serving as a tool for therapists and caregivers of children with ASD; so that, when the child expresses an emotion that can’t be readily identified, the system can recognize it and allow the therapist or caregiver to give the child the necessary tools to help them understand their emotional state. This way they can process the emotion, regulate it, and if necessary, express it in a better way. All with the goal of improving the emotional intelligence of the child. The model was trained with Google’s Teachable Machine software using a dataset that contains images of micro-expressions referencing six basic emotions provided by the Chinese Academy of Sciences Micro-expression (CASME) Databases.