Filtering Motion Vectors using an Adaptive Weight Function

적응적 가중치 함수를 이용한 모션 벡터의 필터링

  • 장석우 (한국건설기술연구원 건설CALS연구센터) ;
  • 김진욱 (한국건설기술연구원 건설CALS연구센터) ;
  • 이근수 (한경대학교 컴퓨터공학과) ;
  • 김계영 (숭실대학교 컴퓨터학부)
  • Published : 2004.11.01

Abstract

In this paper, we propose an approach for extracting and filtering block motion vectors using an adaptive weight function. We first extract motion vectors from a sequence of images by using size-varibale block matching and then process them by adaptive robust estimation to filter out outliers (motion vectors out of concern). The proposed adaptive robust estimation defines a continuous sigmoid weight function. It then adaptively tunes the sigmoid function to its hard-limit as the residual errors between the model and input data are decreased, so that we can effectively separate non-outliers (motion vectors of concern) from outliers with the finally tuned hard-limit of the weight function. The experimental results show that the suggested approach is very effective in filtering block motion vectors.

본 논문에서는 적응적 가중치 함수를 이용하여 블록 단위의 모션 벡터를 필터링하는 방법을 제안한다. 제안하는 방법에서는 먼저 인접한 영상을 받아 들여 가변적 크기의 블록 정합 방법을 이용하여 모션 벡터를 추출한다. 그리고 추출된 모션 벡터를 강건 예측에 적용하여 아웃라이어(outlier)를 제거함으로써 강건 예측에서 사용하는 동작 모델에 근접한 모션 벡터만을 추출한다. 제안된 적응적 강건 예측은 연속적인 시그모이드 가중치 함수를 사용하여 정상 자료와 아웃라이어의 소속 정도를 보다 효과적으로 표현한다. 또한, 최소화 기법의 반복 단계에서 잔여에러가 감소함에 따라 점진적으로 시그모이드 가중치 함수를 조율함으로써 정상 자료와 아웃라이어를 보다 유연하게 분리한다. 실험에서는 카메라의 동작이 포함된 비디오 데이타를 입력 받아 성능을 비교 분석함으로써 제안한 방법의 우수함을 보인다.

Keywords

References

  1. James Davis, 'Mosaics of Scenes with Moving Objects,' Proceedings of Computer Vision and Pattern Recognition, Santa Barbara, CA, pp. 354-360, June 1998. https://doi.org/10.1109/CVPR.1998.698630
  2. Jong-Il Park, Nobuyuki Yagi, Kazumasa Enami, Kiyoharu Aizawa, 'Estimating of Camera Parameters from Image Sequence for Model-Based Video Coding,' IEEE Transactions on Circuits and Systems for Video Technology, Vol. 4, No.3, pp. 288-296, June 1994 https://doi.org/10.1109/76.305873
  3. Eduardo Bayro-Corrochano, Bodo Rosenhahn, 'A Geometric Approach for the Analysis and Computation of the Intrinsic Camera Parameters,' Pattern Recognition, Vol. 35, No. 1, pp. 169-186, January 2002 https://doi.org/10.1016/S0031-3203(00)00182-5
  4. Jong-Il Park, Choong-Woong Lee, 'Robust Estimation of Camera Parameters from Image Sequence for Video Composition,' Signal Processing, Vol.9, pp.43-53, 1996 https://doi.org/10.1016/S0923-5965(96)00009-4
  5. William H. Press, Saul A. Teukolsky, William T. Vetterling, Brian P. Flannery, Numerical Recipes in C : The Art of Scientific Computing, Second Edition, Cambridge University Press, 제15장 : Modeling of Data, pp. 656-706, 1992
  6. Simon Moss, Edwin R. Hancock, 'Registration Incomplete Radar Images Using the EM Algorithm.' Image and Vision Computing, Vol. 15, No. 8. pp. 637-648, August 1997 https://doi.org/10.1016/S0262-8856(97)00014-0
  7. Stan Sclaroff, John Isidore, 'Active Blobs,' Proceedings of International Conference on Computer Vision, Mumbai, India, pp. 1146-1153, 1998 https://doi.org/10.1109/ICCV.1998.710860
  8. B. Liu and A. Zaccarin, 'New fast algorithm for estimation of block motion vectors,' IEEE Trans. Circuits and Systems for Video Tech., Vol.3, pp.148-157, Apr., 1993 https://doi.org/10.1109/76.212720
  9. Reoxiang Li; Bing Zeng Liou, M.L, 'A new three-step search algorithm for block motion estimation Circuits and Systems for Video Technology,' IEEETransactions on, Vol.4, Issue.4, pp.438-442, Aug., 1994 https://doi.org/10.1109/76.313138
  10. Lai-Man Po, Wing-Chung Ma, 'A Novel Four-Step Algorithm for Fast Block Motion Estimation,' IEEE Transactions on Circuits and Systems for Video Technology, Vol. 6, No.3, pp. 313-317, 1996 https://doi.org/10.1109/76.499840
  11. Ramesh Jain, Rangachar Kasturi, Brian G. Schunck, Machine Vision, McGraw-Hill, New York, 1995
  12. Seok-Woo Jang, Hyung-Il Choi, 'A Strategy of Matching Blocks at Multi-Levels,' International Journal of Intelligent Systems, Vol. 17, No. 10, pp. 965-975, 2002 https://doi.org/10.1002/int.10058