Fractional painting training-based optimization enabled deep learning for autism spectrum disorder using MRI images in cloud computing.
Authors
Affiliations (2)
Affiliations (2)
- Research Scholar, Department of Electronics and Communication Engineering, Sri Eshwar College of Engineering, Coimbatore, Tamil Nadu, India. Electronic address: [email protected].
- Professor, Department of Electronics and Communication Engineering, Sri Eshwar College of Engineering, Coimbatore, Tamil Nadu, India.
Abstract
Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder that impairs social interactions. Prior models focus on target individuals older than five years and depend heavily on expert analysis, potentially missing subtle or early indicators of ASD. To address these limitations, Fractional Painting Training-based Optimization-enabled Spiking Attention Neural Network (FPTO_SANN) is devised for ASD. Initially, a cloud is simulated, and brain Magnetic Resonance Imaging (MRI) images of individuals with autism are collected from a specific dataset. Image preprocessing is done by Geometric Mean filtering along with Region of Interest (RoI) extraction, and pivotal regions are identified based on functional connectivity using Fractional Painting Training based Optimization (FPTO). In parallel, feature extraction is conducted on the preprocessed MRI images to obtain Grey Level Co-occurrence Matrix (GLCM) features along with Local Combination Adaptive Ternary Pattern (LCATP) and Discrete Cosine Transform (DCT) features. Finally, ASD classification is performed using a Spiking Attention Neural Network (SANN) that is trained utilizing FPTO, which is developed by integrating Painting Training Based Optimization (PTBO) and Fractional Concept (FC). The FPTO_SANN has attained superior outcomes by considering the metrics, namely accuracy, True Negative Rate (TNR), and True Positive Rate (TPR), with values of 96.67%, 96.98%, and 96.01%.