Uniform Posture Map Algorithm to Generate Natural Motion Transitions in Real-time

자연스러운 실시간 동작 전이 생성을 위한 균등 자세 지도 알고리즘

  • Lee, Bum-Ro (Dept.of Control Instrumentation Engineering, Kwangwoon University) ;
  • Chung, Chin-Hyun (Dept.of Information Control Instrumentation Engineering, Kwangwoon University)
  • 이범로 (광운대학교 제어계측공학과) ;
  • 정진현 (광운대학교 정보제어공학과)
  • Published : 2001.12.01

Abstract

It is important to reuse existing motion capture data for reduction of the animation producing cost as well as efficiency of producing process. Because its motion curve has no control point, however, it is difficult to modify the captured data interactively. The motion transition is a useful method to reuse the existing motion data. It generates a seamless intermediate motion with two short motion sequences. In this paper, Uniform Posture Map (UPM) algorithm is proposed to perform the motion transition. Since the UPM is organized through quantization of various postures with an unsupervised learning algorithm, it places the output neurons with similar posture in adjacent position. Using this property, an intermediate posture of two active postures is generated; the generating posture is used as a key-frame to make an interpolating motion. The UPM algorithm needs much less computational cost, in comparison with other motion transition algorithms. It provides a control parameter; an animator could control the motion simply by adjusting the parameter. These merits of the UPM make an animator to produce the animation interactively. The UPM algorithm prevents from generating an unreal posture in learning phase. It not only makes more realistic motion curves, but also contributes to making more natural motions. The motion transition algorithm proposed in this paper could be applied to the various fields such as real time 3D games, virtual reality applications, web 3D applications, and etc.

동작 포착 시스템에 의해 기록된 동작 데이타를 재사용하는 것은 비용 절감이나 작업과정의 효율성 증대를 위해 매우 중요한 기술이다. 그러나 기록된 데이타의 동작 곡선이 제어점을 가지고 있지 않기 때문에 상호작용을 통한 동작 데이타의 편집이 쉽지 않아서 데이타의 재사용에 어려움이 있다. 기존 동작 데이타를 재사용하는 기술로서 많은 학자들은 조각 동작(Clip motion)들을 부드럽게 연결하여 새로운 동작을 만들어 내는 동작 전이 기술을 제안하고 있다. 본 논문에서는 이러한 동작 전이 기술의 구현 방법으로 균등 자세 지도(Uniform Posture Map: UPM)알고리즘을 제안한다. 학습 단계에서 UPM은 다관절체의 다양한 자세들을 비감독 경쟁 학습을 통해 양자화한다. 이 단계에서 서로 유사한 자세를 나타내는 출력 뉴런을 기하학적으로 근접한 위치에 배치해서 다관절체의 전체 동작 지도를 생성한다. 생성된 UPM의 이러한 특징을 이용해서 적용된 두 동작의 중간 자세를 만들어 내고, 이 자세를 전체 중간 동작을 만들어 내는 키 프레임으로 사용한다. 많은 계산량이 요구되며, 결과 동작을 제어하기가 어려운 다른 동작 전이 알고리즘들과 비교하여 UPM 알고리즘은 중간 동작 생성에 상대적으로 적은 계산량을 요구하며, 하나의 변수를 이용하여 생성된 동작을 제어할 수 있어서 편리한 상호작용 작업 환경을 제공한다. 특히 자기 조직 지도(Self-Organizing Mpa: SOM) 알고리즘을 이용해 자세 지도를 생성할 때, 실제로 존재하지 않은 자세가 포함될 수 있는 가능성을 학습 단계에서 제거함으로써 자세 생성에 있어서 안정성을 확보할 수 있다. 이로 인해 선형 보간법에 비해서 실제 동작에 가까운 동작 곡선을 생성함으로써 보다 자연스러운 동작을 만들어 낼 수 있다. 본 논문에서 제안된 동작 전이 기법은 삼차원 애니메이션 제작, 삼차원 게임, 가상 현실 등의 다양한 분야에 유용하게 적용될 수 있다.

Keywords

References

  1. Soon Ki Jung, Motion Analysis of Articulated Objects For Optical Motion Capture, Ph.D. thesis, Dept. of CS, KAIST, 1997
  2. Douglas J. Wiley and James K. Hahn, Interpolation Synthesis of Articulated Figure Motion, IEEE Computer Graphics and Applications, Vol.7, No.6, pp. 39-45, November, 1997 https://doi.org/10.1109/38.626968
  3. Armin Brudelin and Lance Williams, Motion Signal Processing, Proc. of SIGGRAPH 95, pp.97-104, ACM Press, 1995 https://doi.org/10.1145/218380.218421
  4. Andrew Witkin and Zoran Popovic, Motion Warping, Proc. of SIGGRAPH 95, pp.105-108, ACM Press, 1995 https://doi.org/10.1145/218380.218422
  5. Charles Rose, Bobby Bodenheimer and Michael F. Cohen, Verbs and Adverbs: Multidimensional Motion Interpolation, IEEE Computer Graphics and Applications, v. 18, no 5, pp. 32-40, September. 1998 https://doi.org/10.1109/38.708559
  6. Jehee Lee and Sung Yong Shin, A Hierarchical Approach to Interactive Motion Editing for Human-like Figures, Proc. of SIGGRAPH 99, pp. 39-48, ACM Press, 1999 https://doi.org/10.1145/311535.311539
  7. Michael Gleicher, Motion Editing with Spacetime Constraints, Proc. of Symposium on Interactive 3D Graphics, pp.149-148, 1997 https://doi.org/10.1145/253284.253321
  8. Charles Rose, Brian Guenter, Bobby Bodenheimer and Michael F. Cohen, Efficient generation of motion transitions using spacetime constraints, Proc, of SIGGRAPH 96, pp.147-154, ACM Press, 1996 https://doi.org/10.1145/237170.237229
  9. Matthew Brand and Aaron Hertzmann, Style Machine, Proc of SIGGRAPH 2000, pp.183-192, ACM Press, 2000 https://doi.org/10.1145/344779.344865
  10. Radek Grzeszczuk, Demetri Terzopoulos and Geoffrey Hinton, Neuro.Anirnator: Fast Neural Emulation and Control of Physics- Based Models, Proc. of SIGGHAPH 98, pp.9-20, ACM Press, 1998 https://doi.org/10.1145/280814.280816
  11. Michael F. Cohen, Interactive Spacetime Control for Animation, Proc, of SIGGRAPH 92, pp. 293-302, ACM Press, 1992 https://doi.org/10.1145/133994.134083
  12. Ritter, Martinetz and Schukten. Neural Computation and Self Organizing Maps. Addison Wesley. New York 1992
  13. Stephen I. Gallant, Neural Network Learning and Expert System. MIT Press, 1993
  14. Jianmin Zhao and Norman I.Badler, Inverse Kinematics Positioning Using Nonlinear Programming for Highly Articulated Figures, ACM Transactions on Graphics, Vol.13. No.4, pp.313-336, October, 1994 https://doi.org/10.1145/195826.195827