Autism Spectrum Disorder Detection using Face Features Based on Deep Neural Network

Autism spectrum disorder (ASD) is a neurological condition impacting an individual’s social interaction and communication abilities throughout their life. Classified as a" behavioral variation," ASD manifests typically within the beginning two years of life, although diagnosis may occur at any age. The prevailing theory posits that challenges associated with ASD persist from childhood through adolescence into adulthood. This study explores the potential utility of Convolutional Neural Networks (CNN) and VGG 16 in forecasting and assessing ASD challenges across various age groups, aligning with the growing trend of employing machine learning methodologies in medical diagnostics research. Findings indicate superior performance of CNN-based predictive models across all age categories, achieving heightened accuracy in screening for Autism Spectrum Disorder in the datasets encompassing adults, children, and adolescents. These results were attained subsequent to employing diverse machine learning techniques and addressing missing data concerns.