Rapid generation of custom avatars using depth cameras
Authors: Byoung-Keon Park, Matthew P. Reed
Recent developments in depth - camera technology have enabled these low - cost tools to be used as body scanners. We present a simple software and hardware system using two Microsoft Kinect sensors that can generate an 3D avatar closely matching the body dimensions of an individual in about 15 seconds including scanning, post - processing and modeling time . A custom calibration of the sensors is used to improve measurement accuracy. A statistical body shape model (SBSM) is used to fit the data, thereby overcoming holes, noise, and other limitations. Pilot testing with eight child subjects shows promising results.