Jianhui Zhao
Posture Reconstruction, Hierarchical Structure, Human Animation, Coherence Problem, Criterion Function
This paper presents a hierarchical approach towards human posture reconstruction and animation from monocular video images that contain unrestricted human postures and human movements. A parameterized human skeleton model based on anatomy is used with biomechanical constraints encoded in the joints. Hierarchical structure of the model is studied and applied in Criterion Function to recovery the whole body posture. Angular changes of joints in the same coordinate system between consecutive frames are obtained from hierarchical function of transformation matrixes, and then incorporated into CF to solve the coherence problem in human animation procedure. The experimental results from many monocular images and video sequences are promising.
Important Links:
Go Back
