A fast method for human skeleton tracking is presented, exploiting the human volume reconstructed from 3D reconstruction (Skeleton from Volume – SfV). The method tracks the joint positions of a 15-joints skeletal structure. This structure is separated into a) the rigid-body part that includes the “torso”, “hip”, “neck” and “shoulder” joints, and b) the limb-parts that consist of the “elbow” and “wrist” or the “knee” and “ankle” joints. The proposed method consists of two phases. Initially, in a “user-calibration” phase, the user body structure is estimated. Then, during the “main tracking” phase, both the position-orientation of the rigid-body part and the limb-joint positions are tracked.
The robustness of the method overcomes motion capturing with one Kinect sensor, such as self-occlusions or fast movements. Moreover, motion analysis algorithms allow for repetition detection and exercise evaluation. The human motion capturing and analysis algorithms, integrated into sport and e-health applications, are presented in the video below.