Autism Spectrum Disorder Detection from Structural MRI Using YOLO Deep Learning Models
This study proposes using YOLOv8-11 object detection models on preprocessed structural MRI scans to detect neuroanatomical abnormalities associated with autism spectrum disorder, achieving accuracy rates exceeding 98%. The framework processes 3D MRI data into 2D image sequences across multiple planes while accounting for age-related brain changes, demonstrating superior performance compared to traditional deep learning architectures for potential clinical support in early ASD diagnosis.