DEVELOPMENT OF E-HEALTHCARE SYSTEMFOR AUTISM SPECTRAL DISORDER

Abstract

A neurodevelopmental illness called autism spectrum disorder (ASD) impairs behavior, social interaction, and communication. Effective ASD intervention and therapy depend on early diagnosis.In this project, I'll suggest creating a machine learning algorithm that uses convolutional Early ASD diagnosis using neural networks (CNN) and vocal analysis of children's noises. The proposed algorithm will use audio recordings of children's vocalisations—including screams, coos, babbles, and words—to extract spectral elements like pitch, prosody, and formants that are connected to ASD. The features will be used as input for a CNN that will be trained to distinguish between typical and atypical recordings. We will collect audio recordings from two groups of kids to train and test the algorithm.