Motion capture (MoCap) processing

Abstract

Motion capture technology has become more popular nowadays and widely applied on many applications. Estimating the motion of humans through skeletal tracking gives us another vision to explore and analyse human motion. However, possible viewing occlusions might result in incomplete tracked skeleton by merely using a single camera. This project aims to use two Kinect for Xbox 360 sensors from different angles to capture a more complete human skeleton. Camera calibration is required to combine two skeletons into one complete skeleton. This is done by using one camera as reference for another camera and transform their coordinates into the reference camera’s coordinate system. After transformation, the complete skeleton will be able to be composed.