The Journal of Korean Institute of Communications and Information Sciences (한국통신학회논문지)

The Korean Institute of Commucations and Information Sciences (한국통신학회)
권호 표지

A convergence analysis of a differential method for 2-D motion parameter estimation

미분 추정 기법에 의한 2파원 이동 파라미터 추정의 수렴 특성 분석

  • 이상희 (한국과학기술원 전기 및 전자공학과) ;
  • 유국열 (한국과학기술원 전기 및 전자공학과) ;
  • 김재균 (한국과학기술원 전기 및 전자공학과)
  • Published : 1998.07.01
Download PDF
⟨ Previous Next ⟩

Abstract

In this paper, we investigae convergence behaviors of a differential method for 2-D motion parameter estimation. While the differential method is widely studied for motion compensated prediction in video coding, little attention hs been paid to its convergence properties. Based on the nonseparable exponential covariance image model, we derive the estimates of update terms for the 2- and 6-parameter motion models. And, the effect of noise, spatial correlation, choice of spatial gradient measures, andthe size of a region, are quantitatively anlyzed in relation to the convergence speed. Some empirical results are presented to verify the analysis.

본 논문에서는 2차원 이동 파라미터 추정을 위한 미분 이동 추정 기법의 수렴 특성을 분석한다. 미분 이동 추정 기법은 동영상 부호화에서의 이동 보상 예측을 위하여 많은 연구가 되어 오고 있으나, 그 수렴 특성에 대한 분석은 미미한 실정이다. 본 연구에서는 비분리 지수형 공분산 영상 모델에 근거하여 2-파라미터와 6-파라미터 이동 모델에 대한 파라미터의 추정치를 유도하며, 이를 이용하여, 잡음, 공간, 상관성, 공간 경사 선택 방법, 영역의 크기 등이 수렴속도에 미치는 영향을 정량적으로 분석한다. 분석에 대한 검증을 위하여 몇가지 실험 결과를 보인다.

Keywords

References

  1. IEEE Trans. Comm. v.COM-29 Displacement measur-ement and its application in interframe image coding J. R. Jain;A. K. Jain
  2. Proc. of SPIE Image Coding v.594 A differential displacement estimat-ion algorithm and improved stability M. Bierling
  3. Signal Processing v.16 Differential estimation of the global motion parameters zoom and pan M. Hotter
  4. Signal Processing:Image Communication v.2 no.1 A differential method for simultaneous estimation of rotation, change of scale and translation S. F. Wu;J. Kittler
  5. J. of Visual Comm. Image Rep. v.6 no.4 Robust multire-solutional estimation of parametric motion models J. M. Odobez;P. Bouthemy
  6. Proc. Signal Processing Ⅳ: Theories and Appliations Motion estimation based on a moon model incorporating translation, rotation and zoom G. J. Keesman
  7. Signal Processing v.15 Image segmentation based on object oriented mapping parameter estimation M. Hotter;R. Thoma
  8. IEEE Trans. Circ. Sys. Video Tech. v.5 no.4 Motion compensation using second-order geometric trans-formations C. A. Papadopoulos;T. G. Clakson
  9. ISO/IEC JTC1/SC29/WG11 N1902 Information technology-cod-ing fo audio-visual objects:visuals ISO/IEC
  10. Signal Processing v.35 no.2 An analysis of gradient-based method J.-H. Lee;S.-D. Kim
  11. IEEE Trans. Pattern Anal. and Machine Intell v.PAMI-9 no.2 Optical flow estimation: an error analysis of gradient-based methods with local optimization J. K. Keamey;W. B. Thompson;D. L. Boley
  12. Signal Proces-sing:Image Communication v.6 no.4 On the accuracy and convergence of 2-D motion models using min-imum MSE motion estimitation J.-H. Moon;J.-K. Kim
  13. Proc. IEEE v.69 no.5 Advances in mathematical models for images processing A. K. Jain
  14. Image Sequence Processing and Dynamic Scene Analysis The differential method for image motion estimation C. Cafforio;T. S. Huang (ed.)
  15. Proc. of SPIE Image Coding v.594 Time/space recursions for differential motion estimat-ion R. Robert;C. Cafforio
  16. J.of Visual Comm. Image Rep. v.4 no.1 A gradient-based method for general motion estimation and segmentation S. F. Wu;J. Kittler
  17. Signal Processing v.11 no.4 Motion compensat-ing field interpolation using a hierachically struc-tured displacement estimator M. Bierling;R. Thoma
  18. Signal Processing: Image Communication v.7 no.1 Differential meth-ods for the identification of 2D and 3D motion models in image sequences J.-L. Dugelay;H. Sanson