Improved skeleton estimation by means of depth data fusion from multiple depth cameras

In this work, we address the problem of human skeleton esti-mation when multiple depth cameras are available. We propose a systemthat takes advantage of the knowledge of the camera poses to create acollaborative virtual depth image of the person in the scene which con-sists of points from all the cameras and that represents the person in afrontal pose. This depth image is fed as input to the open-source bodypart detector in the Point Cloud Library. A further contribution of thiswork is the improvement of this detector obtained by introducing twonew components: as a pre-processing, a people detector is applied to re-move the background from the depth map before estimating the skeleton,while an alpha-beta tracking is added as a post-processing step for filter-ing the obtained joint positions over time. The overall system has beenproven to effectively improve the skeleton estimation on two sequencesof people in different poses acquired from two first-generation MicrosoftKinecs.