Autism Spectrum Disorder Detection in Children Using Transfer Learning Techniques

Autism Spectrum Disorder (ASD) is a complex neuro-development disorder that impairs communication, societal interaction and behaviour. Early diagnosis greatly improves the outcomes for individuals with the disorder, as it allows for early intervention and therapy. The traditional methods of diagnosis are often time-consuming, expensive and rely on subjective evaluations. This research work presents a systematic review of existing literature on the use of machine learning for ASD detection and proposes a diagnostic tool, which uses transfer learning techniques to improve the detection of ASD, and classify individuals as either having or not having ASD.