A Skeleton-Occluded Repair Method from Kinect

In this paper, an approach for repairing the occluded human skeleton is proposed. Owing to the perception and record abilities of human beings inherently, people see at a glance that they can intuitively retrieve the moving action for any scene. Human being movement often has a self-occlusions or sheltered phenomenon when they walked. As for this occluded problem, so far, there is not a powerful method to solve, but it is very important scheme especially scene surveillance or patient cares. In order to solve this problem, we propose a skeleton repairing approach for the limbs-sheltered situation. This approach includes human skeleton capture based on the Kinect and natural interaction mechanism, skeleton database construction, and skeleton representation. Experimental results show that the rebuilt human skeleton is very close to the real skeleton like unsheltered state.