We are pretty used to Kinect being used in motion capture of gross motions like that of limbs, but EPFL Computer Graphics and Geometry Laboratory has now also managed to extract fine facial motions from the camera data.
The technology called Faceshift pulls data from the camera and uses a predictive model to match the stream with facial expressions. The company has already released an SDK targeted at animators and game creators and the company is already awaiting the next version of the Kinect, which will allow even finer and more accurate expressions to be captured.
Read more about the technology at Faceshift.com here.
Via The Verge.com