Early Diagnosis of Autistic Children with Eye Tracker and Artificial Intelligence Approach

Abstract

The inadequacies of the traditional methods followed for the diagnosis of autism and the provision of general screening information with non-objective patient relatives observations cause the search for alternative early diagnosis. In addition, constraints on self-expression in young children have increased the need for direct data collection approaches through technological devices and wearable sensors. In particular, using eye tracking sensors to detect eye contact anomalies, which is one of the most important symptoms of autism, can provide important information about risk groups even without other wearable sensors. The first results of our study on the use of eye tracker data for autism diagnosis and processing it with artificial intelligence algorithms are presented here. The results revealed that our algorithm diagnoses with high accuracy, and shows that the eye tracking sensor can be an important tool for early diagnosis of autism.