Interactive Robot-Aided Diagnosis System for Children with Autism Spectrum Disorder

Autism spectrum disorder (ASD) is a group of complex neurodevelopmental disorders characterized by difficulties with social communication and interaction as well as restrictive interest and stereotyped behavior. Despite the behavioral symptoms of ASD often appear early in infancy, the ASD diagnosis is often cumbersome even for expert clinicians owing to characteristic heterogeneity in the symptoms and severity. Early diagnosis and intervention can help children with ASD to achieve more improvement, particularly in their social communication. Here, the study designs an interactive robotic agent and an intelligent image analysis system to assist in the ASD diagnosis of children. The children’s facial expression images and body pose movement images are collected during the human-robot interaction, which three computational models are used for further data analysis. The stored database is presented as a reference for diagnosis in a visual interface. Furthermore, we incorporate multiple AI models in facial emotion recognition and eye tracking detection to automatically analyze images and visualize data, assisting clinicians in diagnostic decision making.