Identifying Autism Spectrum Disorder from Facial Images based on Deep Learning Approach

Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder that is known for causing issues with socio-communication behavior, including anxiety, improper emotional expressions, and sensory sensitivities. Currently, there is no cure for ASD. Early identification, along with proper therapy, can help control behavioral and language development. Traditional methods of ASD detection, which include subjective questionnaires and behavioral observations, are often time-consuming, costly, and require specialized expertise, delaying critical and early inventions. To overcome these challenges, this proposed method provides a novel approach to detecting autism from facial images, utilizing deep learning techniques and the pre-trained convolutional neural network model (VGG16) for feature extraction and classification. The proposed model is validating in real-time scenarios to ensure its effectiveness and practicality in clinical settings by supporting healthcare professionals in delivering timely and effective inventions.