Human Pose Matching Using Skeleton-type Active Shape Models

뼈대-구조 능동형태모델을 이용한 사람의 자세 정합

  • 장창혁 (숭실대학교 미디어학과)
  • Published : 2009.12.15

Abstract

This paper proposes a novel approach for the model-based pose matching of a human body using Active Shape Models. To improve the processing time of model creation and registration, we use a skeleton-type model instead of the conventional silhouette-based models. The skeleton model defines feature information that is used to match the human pose. Images used to make the model are for 600 human bodies, and the model has 17 landmarks which indicate the body junction and key features of a human pose. When applying primary Active Shape Models to the skeleton-type model in the matching process, a problem may occur in the proximal joints of the arm and leg due to the color variations on a human body and the insufficient information for the fore-rear directions of profile normals. This problem is solved by using the background subtraction information of a body region in the input image and adding a 4-directions feature of the profile normal in the proximal parts of the arm and leg. In the matching process, the maximum iteration is less than 30 times. As a result, the execution time is quite fast, and was observed to be less than 0.03 sec in an experiment.

본 논문은 뼈대-구조(skeleton) 형태의 Active Shape Models을 이용한 사람의 자세 정합에 대한 새로운 접근 방법을 제안한다. 제안된 방법은 모델 생성과 정합 과정에서의 빠른 수행 시간을 위해 기존 윤곽 형태(silhouette)의 모델이 아닌 뼈대-구조 형태의 모델을 적용하였다. 기존 Active Shape Models을 뼈대-구조 형태로 사람 자세 정합에 적용했을 경우 자세를 결정짓는 팔과 다리의 부정확한 정합은 사람 몸의 다양한 색상 정보와 전후(fore-rear direction)만을 고려한 특징점(landmark)의 방향정보로 인해 발생되며, 이러한 문제점은 입력 영상의 차영상 정보와 사람의 자세를 결정짓는 팔과 다리의 중요 특징점에 방향정보를 추가하여 해결하였다. 사람의 뼈대-구조 모델을 생성하기 위해 600개의 이미지를 사용 하였으며, 생성된 형태 모델은 사람의 자세에 정합될 수 있는 17개의 특징점을 포함한다. 정합 과정에서 최대 30번 이하의 반복 과정을 수행 하며, 최대 수행 시간은 0.03초로 빠른 수행 시간의 결과를 얻었다.

Keywords

References

  1. R. Poppe, 'Vision-based human pose analysis: An overview,' Computer Vision and Image Understanding, vol.108, pp.4-18, 2007 https://doi.org/10.1016/j.cviu.2006.10.016
  2. M. Brand, 'Shadow puppetry,' Proceedings of the International Conference on Computer Vision (ICCV99), vol.2, pp.1237-1244, 1999 https://doi.org/10.1109/ICCV.1999.790422
  3. A. B. Albu, R. Bergevin, and S. Quirion, 'Generic temporal segmentation of cyclic human pose,' Pattern Recognition, vol.41, pp.6-21, 2008 https://doi.org/10.1016/j.patcog.2007.03.013
  4. T. F. Cootes, C. J. Taylor, 'Statistical Models of Appearance for Computer Vision,' http://personalpages. manchester.ac.uk/staff/timothy.f.cootes/Models/app_models.pdf
  5. M. Kass, A. Witkin, and D. Terzopoulos, 'Snakes: Active contour models,' In 1st International Conference on Computer Vision, pp.259-268, London, June 1987
  6. R. Navaratnam, A. Thayananthan, P. H. Torr, and R. Cipolla, 'Hierarchical part-based human body pose estimation,' Porceedings of the British Machine Vision Conference(BMVC'05), Oxford, United Kingdom, pp.479-488, 2005
  7. X. Shanon, J. B. Michael, and Y. Yacoob, 'Cardboard people: A parameterized model of articulated image pose,' Proceedings of the International Conference on Automatic Face and Gesture Recognition(FGR 96), Killington, VT, pp.38-44, 1996
  8. I. Haritaoglu, D. Harwood, and L. S. Davis, 'W4s: A realtime system detecting and tracking people in 2 1/2D,' Proceedings of the European Conference on Computer Vision(ECCV'98), LNCS, vol.1(1406), pp.877-892, 1998
  9. N. R. Howe, M. E. Leventon, and W. T. Freeman, 'Bayesian reconstruction of 3D human pose from single-camera video,' Advances in Neural Information Processing Systems(NIPS) 12, Denver, CO, pp.820-826, 2000
  10. D. Kim, V. Maik, D. Lee, J. Shin, and J. Paik, 'Active Shape model-based object tracking in panoramic Video,' Proc. ICCS, LNCS, vol.3994, pp.922-929, 2006
  11. A. M. Baumberg, and D. C. Hogg, 'An Efficient Method for Contour Tracking using ASM,' Pose of Non-Rigid and Articulated Objects, Proceedings of the 1994 IEEE, pp.194-199
  12. C. H. Jang, and K. C. Jung, 'Human pose estimation using ASM,' Proceedings of World Academy of Science, Engineering and Technology (WASET), vol.34, pp.312-316, 2008
  13. D. Shi, S. R. Gumm, and R. I. Damper, 'Handwritten Chinese Radical Recognition Using Nonlinear ASM,' IEEE Transactions on Pattern Analysis and Machine Intelligence, vol.25, no.2, pp.277-280, 2003 https://doi.org/10.1109/TPAMI.2003.1177158
  14. T. F. Cootes, C. J. Taylor, D. H. Cooper, and J. Graham, 'ASM-Their Training and Application,' Computer Vision and Image Understanding, vol. 61, pp.38-59, 1995 https://doi.org/10.1006/cviu.1995.1004
  15. I-C. Chang, and C.-L. Huang, 'Skeleton-based Walking Pose Analysis Using Hidden Markov Model and ASM,' Journal of Information Science and Engineering, vol.17, pp.371-403, 2001
  16. H. Sunder, D. Silver, N. Gagvani, and S. Dickinson, 'Skeleton Based Shape Matching and Retrieval,' Shape Modeling International, pp.130-139, 2003
  17. H. Blum, 'A Transformation for Extracting New Descriptors of Shape,' Computer Vision and Image Understanding: MIT press, pp.362-380, 1967
  18. B. V. Ginneken, A. F. Frangi, J. J. Staal, B. H. Romeny, and M. A. Viergever, 'Active Shape Model Segmentation With Optimal Features,' IEEE Transactions on Medical Imaging, vol.21, no.8, pp.924-933, 2002 https://doi.org/10.1109/TMI.2002.803121
  19. H. H. Thodberg, and A. Rosholm, 'Application of the active shape model in a commercial medical device for bone densitometry,' Image and Vision Computing, vol.21, pp.1155-1161, 2003 https://doi.org/10.1016/j.imavis.2003.09.002
  20. T. F. Cootes, A. Hill, C. T. Taylor, and J. Haslam, 'The use of ASM For Locating Structures in Medical Images,' Image and Vision Computing, vol.12, no.6, pp.355-366, 1994 https://doi.org/10.1016/0262-8856(94)90060-4
  21. W. Wang, S. Shan, W. Gao, B. Cao, and B. Yin, 'An Improved Active Shape Model for Face Alignment,' Proceedings of the Fourth IEEE International Conference on Multimodal Interfaces, pp.523-528, 2002
  22. S. Romdhani, S. Gong, and A. Psarrou, 'A Multi-View Nonlinear Active Shape Model Using Kernel PCA,' (BMVC'99), pp.483-492