While children play at the Shirley G. Moore Laboratory School their every move is being recorded by five Kinect motion sensors tucked away in the corners of the room. No, this isn't some clever new security system or some Orwellian plot by school administrators; the Microsoft motion sensing game technology for Xbox 360 and PC is being used to detect autism spectrum disorder (ASD) in young children. The New Scientist offers an extensive report on the program (uncovered by Eurogamer).
This set-up was put together by researchers from the University of Minnesota's Institute of Child Development in Minneapolis and it is designed to look for signs of behavioral disorders. The goal of this project is to figure out if Microsoft's game sensor will work with computer-vision algorithms trained to detect behavioral abnormalities to automate the process of diagnosing early signs of autism.
The program was put together by Guillermo Sapiro, Nikolaos Papanikolopoulos and other colleagues, who fitted the nursery with five Kinects to monitor groups of children (ten or more) ages 3 – 5 as they played.
The Kinects identify and track children based on their shape and the color of the clothes they are wearing, and the information on their movement is then fed to three PCs running software that logs each child's activity level – including limb movements. The computer then compares the data against the room's average. Researchers say the system can flag children who are hyperactive or unusually still – which can be possible signs of autism.
"The idea is not that we are going to replace the diagnosis, but we are going to bring diagnosis to everybody," Sapiro says. "The same way a good teacher flags a problem child, the system will do automatic flagging and say, 'Hey, this kid needs to see an expert'."
The team's ultimate goal is to combine the Kinect-enabled detection system with another project it is working on: computer-vision algorithms that can identify behavioral markers designated on the Autism Observation Scale for Infants. The system measures traits such as the ability to follow an object when it passes in front of the eyes, certain mannerisms and posture that are considered early signs of a possible ASD.
"Early diagnosis is critical in helping people with autism get the support they need," says Caroline Hattersley of The National Autistic Society in London. She stresses that autism specialists are still needed. "While this technology could potentially identify some signs of autism, there are many factors, such as language delay and limited eye contact that it would miss."
"We are trying to do very difficult and expert analysis that a psychiatrist would do, but automatically," says Sapiro. He envisions a future video game for Kinect that could test a child as they played with a parent, for example.
The system will be presented at the IEEE International Conference on Robotics and Automation in St Paul, Minnesota later this month.