Autism spectrum disorder (ASD) is a world-threatening mental developing disorders that recently appeared widely, due to its diagnosis complexity as well as lack of evidence of its real causes. Many researchers have afforded great effort to precisely identify this syndrome and its symptoms. This survey provides a comprehensive study of autism spectrum disorder, its types, symptoms, prevalence, and developments in its diagnosing. Six categories for autism exposure and identification are currently investigated; clinical monitoring, genetics and blood analysis, Functional magnetic resonance imaging (fMRI), Electroencephalography (EEG) based investigation, wearable sensors and finally computer vision-based techniques. Computational technologies, especially computer-vision, machine learning and neural networks techniques have added great advances in detecting autism and these techniques are comprehensively reviewed in this paper. Also, medical assisting computer vision-based framework is proposed to detect observable autism symptoms. The proposed framework utilises recent and efficient techniques that can be used to produce accurate diagnosing results.