A technology that constantly watches over children could help identify the revealing signs of autism, thereby prompting an early diagnosis for the same. University of Minnesota's Institute of Child Development in Minneapolis boasts of five Kinect motion sensors at the corners of the room, recording each and every move of each and every student under them.
A typical setup has been developed to observe any signs of behavioral disorders. The aim is to check whether Kinect's sensor, plus the computer-vision algorithms directed to identify these symptoms, could be employed to automate the prompt detection of autism. Discovering an Autism Spectrum Disorder (ASD) is quite difficult in the case of young children, but it's preferred to turf the safer side by introducing the child to speech therapy and socio-communication skills. A lot many symptoms point towards ASD, but again, they are all quite faint. An experienced diagnostician may very well detect the signs after thoroughly going through the video footage of the child engaged in physical activities, but this is quite time-consuming.
The officials at the nursery believe that fitting 5 cameras in the classroom for the surveillance of 10 students aged 3-5, should help them automate this process. The cameras shall pick out and track the children based on their shapes and the color of clothes they have on them. The system shall be presented before the IEEE International Conference on Robotics and Automation committee in St. Paul, Minnesota this very month.