Using Gaze Tracking on Sensory Integration Therapy for Eye-movement Detection

Abstract

Autism spectrum disorder is a condition that often occurs in infancy (from birth to three years old). A common problem includes eyesight disorders, where children cannot focus on a specific item. We use the method of Convolutional Neural Networks (CNN) to detect eyes and collect eye positions. Then, we use a linear regression model to best fit the movement of the eyeballs through the track of positions. After that, we use an Analytic Hierarchy Process (AHP) model for a multi-perspective judgment on children's performance in eye movement.