Autism Spectrum Disorder Detection Using Fractional Social Driving Training-Based Optimization Enabled Deep Learning.

Autism Spectrum Disorder (ASD) is neurodevelopment-based impact on interactive communication and social skills. Diagnosing ASD is one of serious issues that start manifesting at low ages, and is difficult to diagnose at early stages. Autism is characterized by both environmental and genetic factors. Lack of communication issues, social interaction, and limited interest behaviors are possible individuality of autism noticed in children, along other symptoms. This paper aims at ASD detection by Deep Quantum Neural Network (DQNN), wherein this network is trained by proposed Fractional Social Driving Training-Based Optimization (FSDTBO). The initial stage of this processing starts with acquisition of image from dataset, and further pre-processing is carried out using Gaussian filter, and this filtered image is suspended for Regions of Interest (ROI) extraction. Also, extraction of nub region is done by proposed Social Driving Training-Based Optimization (SDTBO), from which classification process is done by considering extracted features too. Here, proposed FSDTBO is integration process among Fractional Calculus (FC) and SDTBO, wherein SDTBO is collaboration between Social Optimization Algorithm (SOA) and Driving Training-Based Optimization (DTBO). Moreover, classification performance of ASD is found based on three metrics, like accuracy, specificity, and sensitivity with superior values of 0.90, 0.94, and 0.96. [ABSTRACT FROM AUTHOR]