Measurement of Line-of-Sight Detection Using Pixel Quantity Variation and Application for Autism

Abstract

In this study, we develop a tool to support physically disabled people's communication and an assessment tool to measure the intelligence index of autistic children, which uses eye movements with image processing. For the measurement of eye movements, we newly developed a pixel center of gravity method that detects in which the direction of the eye movement is shown in the point where the weights of the black pixels moved. This method is different from using the conventional black eye detection or ellipse detection. The method enables accurate detection even when a physically handicapped person uses. On the other hand, the assessment tool that measures the intelligence index of autistic children prepares dedicated goggles that combines light emitting diodes and near-infrared cameras. It is a study to measure the response speed of left and right eye movements by applying the results so far and to explore the relationship with autism.