Prediction of Autism Spectrum Disorder using Convolution Neural Network

Autism Spectrum Disorder is a developing neurological disorder, and it starts in childhood, adolescence, and adulthood. The treatment of Autism Spectrum Disorder (ASD) necessitates an accurate diagnosis followed by appropriate rehabilitation. Physicians can use artificial intelligence (AI) technology to help them implement computerized diagnosis and rehabilitation processes. Neuroimaging-based approaches have been the focus of deep learning algorithms for ASD diagnosis. Neuroimaging techniques are benign disease indicators that could aid in identifying ASD. Neuroimaging procedures, both structural and functional, give doctors much information about the brain's anatomy and activity. Because of the brain's complex structure and function, developing optimal processes for ASD identification using neuroimaging data deprived of using Deep Learning is difficult. Our proposed work aims to identify Autism Spectrum Disorder(ASD) from a huge dataset based on brain patterns. Using a convolution neural network, the proposed work identifies ASD patients from ordinary people. It identifies the ROI using the feature extraction technique. The system performance is measured by accuracy, and it achieves 95% accuracy to identify ASD patients.