Identification of Autism Spectrum Disorder (ASD) using Autoencoder

Abstract

Deep Learning (DL) techniques are computational models based on representation learnings. They are demonstrated to be the best reasonable strategies to deal with information with various portrayals and with numerous degrees of reflection. Recognizable proof of ASD has been a test as there is no demonstrated reason for it. The issue has been tended to by numerous specialists with the utilization of fMRI. As MRI and its varieties have 3D representations, Machine Learning and Deep Learning techniques are appropriate to deal with and handle them. This paper extends the recognizable proof of ASD from fMRI pictures utilizing Autoencoder organize. The examinations are led on the benchmark dataset ABIDE II. Results uncover that DL strategies are bringing out better classifiers delivering a great degree of arrangement exactness.