An Approach to Detect Autism in Early Stage Using Inception v3

The introduction of artificial intelligence has revolutionized multiple domains, including the medical sector. AI has played a major role in assisting people with complex tasks. One such advancement is computer vision, a deep learning technology that helps the computer visualize things and, therefore, helps in tasks like recognition and classification. There are certain advancements in the medical sector as AI helps in identifying certain issues. One such advancement is that the world has encountered a rapid increase in the number of people who are autistic, not because many people have developed autistic traits but due to the advancements in tools and methods used to detect autism and awareness about autism. Autism Spectrum Disorder is a neuro-developmental disorder characterized by repetitive behaviour, difficulty in social interaction and communication, and specific and odd interests. Existing systems to detect autism include clinical diagnosis and genetic testing to look for genetic markers that are linked to autism. Studies are being carried out to assist in detecting autism by analysing MRIs, understanding brain patterns, and AI-based behavioural analysis. Our solution to detect autism uses a convolutional neural network which is a deep learning technique that uses facial images to detect autism by utilizing the Inception v3 model. The main focus of this study relies on finding a suitable solution that assists in detecting autism at the earliest, as many developing countries have lower reported cases of autism due to underdiagnosis and lacking awareness.