XceptionNet-based Deep Learning Framework for Accurate and Efficient Autism Spectrum Disorder Detection

Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder characterized by complex cognitive, social, and behavioral functioning, for which early and accurate diagnosis is mandatory to implement effective intervention. Most traditional approaches to its diagnosis rely on behavioral assessments that are subjective and very time-consuming, hence delayed detection. Recent development in the field of deep learning and medical applications has paved the way toward automated ASD classification; however, current models face issues of efficiency in feature extraction, computational complexity, and accuracy in classification. In consideration of these challenges, this paper proposes an ASD detection system using XceptionNet with depthwise separable convolution and an optimized feature extraction strategy for better classification accuracy with reduced computational overhead. The proposed model is compared to the state-of-the-art deep learning architectures, and it results in better performance in terms of accuracy, precision, recall, and F1-score. The final proposed XceptionNet-based ASD detection system is the high-accuracy, computationally efficient solution for automated ASD classification, overcoming several major limitations of previous approaches.