Utilizing Hybrid-Deep Learning for Autism Spectrum Disorder Detection in Children via Facial Emotion Recognition

Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder that significantly impacts social interaction and communication. Early detection and intervention are crucial for optimal outcomes. This research investigates the use of deep learning techniques to identify ASD based on facial emotion analysis. The proposed approach leverages Self-Attention-Based Progressive Generative Adversarial Networks (SA-PGAN) optimized with the Gorilla Troops Optimizer (GTO) to accurately recognize subtle facial cues associated with ASD. The model is trained on a dataset of facial images from autistic children, undergoing extensive preprocessing to ensure data quality and consistency. The performance of the proposed model is evaluated using various metrics, including accuracy, precision, recall, F1-score, and computational time. The results demonstrate significant improvements over state-of-the-art methods, such as Random Forest (RF), Support Vector Machine (SVM), and Convolutional Neural Networks (CNN). This research contributes to the development of advanced AI-powered tools for early ASD detection, enabling timely interventions and improving the quality of life for individuals with ASD.