Towards Model-free Markerless Motion Capture

摘要

An approach for model-free markerless motion capture of humans is presented. This approach is centered on generating underlying nonlinear axes (or a skeleton curve) from a volume of a human subject. Human volumes are captured from multiple calibrated cameras. We describe the use of skeleton curves for determining the kinematic posture of a human captured volume. Our motion capture uses a skeleton curve, found in each frame of a volume sequence, to automatically produce kinematic motion. We apply several types of motion to our capture approach and use the results to actuate a dynamically simulated humanoid robot.