Detecting Emotions in Autism Spectrum Disorder via Deep Learning Algorithms
by Minju Kang
Abstract – The prevalence of Autism Spectrum Disorder (ASD) has been increasing in recent years, and early diagnosis, intervention are crucial for patient outcomes. Artificial Intelligence (AI) has shown potential in detecting ASD through machine learning algorithms. This study aimed to use deep learning algorithms to detect ASD and the emotional state of ASD patients through facial expressions. The study found that the MobileNetV2 algorithm performed the best, achieving an accuracy of 80% in detecting emotions of ASD patients. Continued research in this area may lead to earlier, better diagnosis of autism and improve the quality of care for individuals with ASD.