An Intelligent System for Real-Time Monitoring of Autistic Children

Abstract

Children with Autism Spectrum Disorder (ASD) face communication and social interaction difficulties, often exhibiting stereotyped behaviors like hand-flapping or head rocking, and struggle to maintain concentration. These issues hinder their integration and make manual evaluation subjective and limited. In this paper, we developed an AI-based intelligent system that uses computer vision to track and analyze these behaviors in real-time. For attention estimation, the KNN classifier reached an overall accuracy of 98.6% with an F1-score of 98%. This approach provides an effective and non-invasive tool to support therapists and parents in monitoring autistic children and offers a promising step towards automated behavioral assessment.