Emotion detection in children with autism spectrum disorder: leveraging EEG signals and companion bots for enhanced interaction

Children with Autism Spectrum Disorder (ASD) often struggle with emotional recognition and expression, which can hinder their social interactions and communication skills. Detecting emotions in children with Autism Spectrum Disorder is crucial for understanding their emotional states, as these children often face challenges in recognizing and expressing feelings. This paper explores the facial emotions and EEG signals for emotion detection with machine learning, deep learning, and physiological signal. The results show substantial accuracy in recognizing emotions by utilizing multimodal data and powerful algorithms, providing insightful information for educators and caregivers. Our goal in modifying YOLO for emotion identification is to take use of its accuracy and speed in dynamic contexts like real-time video streams. The study assesses how effectively the YOLO architectures—more especially, YOLOv5s, YOLOv7-tiny, and YOLOv8s—perform on a customized dataset intended for the identification of emotions in kids with ASD. The results demonstrate the models' efficacy in precisely recognizing emotional expressions, which supports therapeutic interventions for kids with ASD.