Research Article
Automatic and Interactive Key Posture Design by Combing the PIK with Parametric Posture Splicing
@INPROCEEDINGS{10.1007/978-3-642-02315-6_12, author={Shilei Li and Bing Wu and Jiahong Liang and Jiongming Su}, title={Automatic and Interactive Key Posture Design by Combing the PIK with Parametric Posture Splicing}, proceedings={Intelligent Technologies for Interactive Entertainment. Third International Conference, INTETAIN 2009, Amsterdam, The Netherlands, June 22-24, 2009. Proceedings}, proceedings_a={INTETAIN}, year={2012}, month={5}, keywords={character animation posture designing prioritized inverse kinematics parametric posture splicing}, doi={10.1007/978-3-642-02315-6_12} }
- Shilei Li
Bing Wu
Jiahong Liang
Jiongming Su
Year: 2012
Automatic and Interactive Key Posture Design by Combing the PIK with Parametric Posture Splicing
INTETAIN
Springer
DOI: 10.1007/978-3-642-02315-6_12
Abstract
Key posture design is commonly needed in computer animation. This paper presents an automatic and interactive whole body posture designing technique by combining the PIK (prioritized inverse kinematics) with the proposed parametric human posture splicing technique. The key feature of PIK is that the user can design a posture by adding high level constraints with different priorities. However, the PIK is essentially a numerical IK algorithm which relies on the iterative optimization starting from a good enough initial posture to get the final result. To speed up the running efficiency and ensure the lifelikeness of the final posture, the parametric posture splicing technique is proposed to generate the initial guess of the PIK. According to the set of the high level constraints, the whole body is divided into some partial parts, whose postures are then generated by the parametric posture synthesis from a single posture database. Then an initial posture guess with some main characteristics of the finally acceptable posture can be generated approximately by splicing these partial body postures together. Starting from this initial guess and with all constraints considered at different priority levels, the PIK can be initialized with a bias defined by this particularly initial guess and iterated step by step to get a final posture. The total process of the whole body posture generation is automatic and interactive. The experimental results show that this combination method can not only improve the computation efficiency of the PIK but also can simultaneously ensure the naturalness of the final posture.