Varied computational intelligent techniques for early identification and exploration of autism spectrum disorder in children

Abstract

Autism Spectrum Disorder (ASD) is one of the predominantly exploring neurodevelopment disorders that seek early identification and exploration in children. As Autism Society of India has analysed, it is observed that 1 in every 168 children are diagnosed with autism in India. Owing to this rapid increase in the rate of people with high functioning ASD, many researchers and medical organizations have joined hands to identify the underlying factors behind the disorder. The basic idea behind the evaluation of autistic characteristics is to impart high quality life style to the children with ASD that matches the regular developing toddlers. The screening and diagnosis of autism could be made through analysis of the social behaviour of the child, eye gaze factor, linguistic ability, responsiveness, creativity, etc in their various stages of development. These abilities of the child are analysed through regular dyad communication made towards the child by the parents, caretakers and clinicians. These inferences of the dyads are profiled for clinical evaluation and analysis to identify the level of autism in the child. The proposed research aims in bridging the gap identified between the early identification biomarker and the existing diagnostic procedures. As the responsiveness and the behaviour of any human could be perceived under the expression publicized in their face, analysis of facial expression is considered the highest priority. The facial expression exposed by the children with ASD and TD are analyzed using SVM Linear Kernel Classifier that classifies the expression into and#8213;angerand#8214;, and#8213;happyand#8214;, and#8213;fearand#8214;, and#8213;sadand#8214;, and#8213;disgustand#8214;, and#8213;neutraland#8214; and and#8213;sleepand#8214;. newline