3D Eye Gaze tracking to Estimate Joint Attention
Sisil Mehta, Gregory Abowd, Irfan Essa
Joint attention is a distinguishing characteristic of children with Autism Spectrum Disorders(ASD) and is used for diagnosis of ASD in various screening and diagnosis methodologies (MCHAT, ADOS). However using computational techniques for estimating instances of Joint Attention can complement current diagnosis methods – aid less experienced clinicians or act as a warning tool for parents in countries where autism screening is not carried out on a per child basis. We propose a low cost solution for estimating instances of Joint Attention using an interactive task such as ball play between a clinician and a child. We overcome the difficulties associated with conducting trial experiments with children by making the tasks engaging. The larger objective is to provide large scale objective screening techniques to identify children at risk so that timely intervention therapies are made available to them.