Leveraging Machine Learning Techniques for Detecting Emotional States in Asd Children

Abstract

ASD mainly correlates with brain development with neuro illness which impacts the overall perception of people mental stability. Everyone comes across different situations every day and experiences different emotions. A characteristic person can express themselves in writing or verbally. Children with autism, on the other hand, find it difficult to put their feelings into words. This happens because the intellect has not matured to a characteristically human level. People feel uncomfortable expressing their feelings in the same way that normal people do. Without acknowledgement and reassurance, children can react violently, with serious consequences. Their sentiments are usually either happy, sad, angry, or indignant. To fully conceive their situation, we present how collaborative machine learning techniques can be used to analyse the situation and predict their emotions. Capture their expressions at all times or amidst unusual activity. This classifier model aids in predicting an autistic child's emotions at each stage. We can examine images and extract facial features. Intensity can be estimated using the obtained properties. Based on these, we can predict the emotions of such a person. This aim of therapy is to enhance the children's functioning by minimising the indications of ASD and fostering growth and education. This research acknowledges the Kaggle dataset for image classification to determine facial expressions. The ASD classification increasingly includes many disorders that were primarily classified as separate disorders.