An interactive facial-expression training platform for individuals with autism spectrum disorder

摘要

Emotions are an incredibly important aspect of human communication. Individuals with Autism Spectrum Disorder (ASD) suffer from significant challenges in perceiving, and understanding others' emotions, and responding emotionally in an appropriate manner. In order to help them to improve their facial expression recognition and response skills, EmoTrain was developed as an interactive platform that utilizes deep learning to track a user's facial expressions from a mobile device's camera and assists a user in learning to understand facial expressions and how to perform them. The paper details the design of the EmoTrain platform, the deep learning model and game engine for the platform, and evaluates EmoTrain's effectiveness.