A Machine Learning Approach for Early Detection and Diagnosis of Autism and Normal Controls and Estimating Severity Levels Based on Face Recognition

Autism is unique among the numerous brain disorders in that it typically affects children at a young age. For people with autism, the most difficult element is expressing their sentiments and emotions to others. Autism spectrum disorder(ASD) is another name for autism is a chronic developmental impairment, difficult and complex, marked by recurring actions, non-verbal communication, and lack of concentration. Although autism cannot be cured, early diagnosis can assist in reducing its symptoms. ASDs have varying degrees of symptoms and severity. This study uses the most well-known machine learning techniques to discriminate between autistic people and healthy controls. Thermal-face photos are used to extract features using GLCM. This was accomplished using autocorrelations, contrast, cluster, prominence, cluster shadow, difference, entropy, squared sum variance, homogeneity, and the maximum probability. For example, the Support-Vector Machine Classifier, K-Nearest Neighbor Classifier, Nave Bayes Classifier, and Random-Forest Classifier have been utilised for classification.