Title

使用核關聯之非剛性形體對齊與對應

Translated Titles

Non-Rigid Shape Registration Using Kernel Correlation

DOI

10.6842/NCTU.2013.00806

Authors

林育右

Key Words

形體對齊與對應 ; 核關聯 ; Shape Registration ; Kernel Correlation

PublicationName

交通大學資訊科學與工程研究所學位論文

Volume or Term/Year and Month of Publication

2013年

Academic Degree Category

碩士

Advisor

莊榮宏;黃世強

Content Language

英文

Chinese Abstract

我們提出一個非剛性形體間之對齊與對應的方法。許多基於形變的方法改進迭代最近點法,並且將形體之對齊與對應的問題轉化為最佳化問題。然而,基於迭代最近點法的非線性能量系統必須在每次迭代時改變對應的最近點,並且移除當中被認為是不好的對應。此項行為造成最佳化的過程中改變了能量系統,因此最佳化處理時無法直截了當地求解。相反地,我們使用基於核關聯的方法來表示非線性能量系統,此法間接地給了動點一個方向,並且保證最佳化過程中擁有固定的數學表示式。我們的演算法不將每個形變物表上的取樣點對應到目標表物表上的另一個取樣點,而是將其對應到目標物表上的一個合理位置,因此獲得更合意的對齊與對應結果。此外,我們的演算法因擁有固定的數學表示式而能更有效率地求出最佳解。

English Abstract

We present an algorithm for shape registration of non-rigid partial scans. Many deformation-based methods adapt their algorithm based on iterative closest point (ICP) and formulate the registration as an optimization problem. However, non-linear energy systems based on ICP should change the set of closest points iteratively and remove some of them for filtering out bad correspondence. This behavior changes the formulation during energy minimization and the optimization process can not be solved straightforwardly. On the contrary, we formulate the energy system using kernel correlation (KC), which implicitly gives a direction for a moving point and guarantees a fixed formulation during optimization. Our algorithm gains a more preferable result since each point on the source surface is not fitted to a point but a reasonable 3D position on the target surface, and can be more efficient due to the fixed formulation of energy equations.

Topic Category 基礎與應用科學 > 資訊科學
資訊學院 > 資訊科學與工程研究所
Reference
  1. body shapes: reconstruction and parameterization from range
    連結:
  2. for robust pairwise surface registration. ACM Transactions on
    連結:
  3. to ne surface registration without initial motion estimation.
    連結:
  4. In IEEE Conference on Computer Vision and Pattern
    連結:
  5. [BM92] P. J. Besl and N. D. McKay. A method for registration of 3-D
    連結:
  6. shapes. IEEE Transactions on Pattern Analysis and Machine
    連結:
  7. Intelligence, 1992.
    連結:
  8. symposium on Geometry processing, 2005.
    連結:
  9. [HCTW11] H. Huang, J. Chai, X. Tong, and H.-T. Wu. Leveraging motion
    連結:
  10. capture and 3d scanning for high-delity facial performance acquisition.
    連結:
  11. geometry and motion reconstruction. ACM Transactions on
    連結:
  12. S. Rusinkiewicz. Temporally coherent completion of dynamic
    連結:
  13. of the Symposium on Geometry Processing, 2008.
    連結:
  14. least squares problems. Technical report, Informatics
    連結:
  15. and Mathematical Modelling, Technical University of Denmark,
    連結:
  16. Symposium on Advances in Visual Computing, 2009.
    連結:
  17. graphics Symposium on Geometry Processing, 2011.
    連結:
  18. Globally consistent space-time reconstruction. In Proceed-
    連結:
  19. and H.-P. Seidel. Animation cartography - intrinsic reconstruction
    連結:
  20. of shape and motion. ACM Transactions on Graphics -
    連結:
  21. [TK04] Y. Tsin and T. Kanade. A correlation-based approach to robust
    連結:
  22. Vision, 2004.
    連結:
  23. Sampled Vision Problems. PhD thesis, School of Computer Science,
    連結:
  24. P. Jenke, L. Guibas, H.-P. Seidel, and A. Schilling. Ecient
    連結:
  25. reconstruction of nonrigid shape and motion from real-time 3d
    連結:
  26. [WJH+07] M. Wand, P. Jenke, Q. Huang, M. Bokeloh, L. Guibas, and
    連結:
  27. A. Schilling. Reconstruction of deforming geometry from timevarying
    連結:
  28. symposium on Geometry processing, 2007.
    連結:
  29. [ACP03] B. Allen, B. Curless, and Z. Popovic. The space of human
  30. scans. ACM Transactions on Graphics - Proceedings of ACM
  31. SIGGRAPH, 22(3), 2003.
  32. [AMCO08] D. Aiger, N. J. Mitra, and D. Cohen-Or. 4-points congruent sets
  33. Graphics - Proceedings of ACM SIGGRAPH, 27(3), 2008.
  34. [ART10] A. Albarelli, E. Rodola, and A. Torsello. A game-theoretic approach
  35. Recognition, 2010.
  36. [BHB+11] T. Beeler, F. Hahn, D. Bradley, B. Bickel, P. Beardsley, C. Gotsman,
  37. R. W. Sumner, and M. Gross. High-quality passive facial
  38. performance capture using anchor frames. ACM Transactions
  39. on Graphics - Proceedings of ACM SIGGRAPH, 30(4), 2011.
  40. [GMGP05] N. Gelfand, N. J. Mitra, L. J. Guibas, and H. Pottmann. Robust
  41. global registration. In Proceedings of the third Eurographics
  42. ACM Transactions on Graphics - Proceedings of ACM
  43. SIGGRAPH, 30(4), 2011.
  44. [LAGP09] H. Li, B. Adams, L. J. Guibas, and M. Pauly. Robust singleview
  45. Graphics - Proceedings of ACM SIGGRAPH Asia, 28(5), 2009.
  46. [LLV+12] H. Li, L. Luo, D. Vlasic, P. Peers, J. Popovicand M. Pauly, and
  47. shapes. ACM Transactions on Graphics - Proceedings of ACM
  48. SIGGRAPH, 31(1), 2012.
  49. [LSP08] H. Li, R. W. Sumner, and M. Pauly. Global correspondence optimization
  50. for non-rigid registration of depth scans. In Proceedings
  51. [MNT04] K. Madsen, H. B. Nielsen, and O. Tingle. Methods for nonlinear
  52. 2004.
  53. [PB09] C. Papazov and D. Burschka. Stochastic optimization for rigid
  54. point set registration. In Proceedings of the 5th International
  55. [PB11] C. Papazov and D. Burschka. Deformable 3d shape registration
  56. based on local similarity transforms. In Proceedings of Euro-
  57. [PSDB+10] T. Popa, I. South-Dickinson, D. Bradley, A. Sheer, and W. Heidrich.
  58. ings of Eurographics Symposium on Geometry Processing, 2010.
  59. [SSP07] R. W. Sumner, J. Schmid, and M. Pauly. Embedded deformation
  60. for shape manipulation. ACM Transactions on Graphics -
  61. Proceedings of ACM SIGGRAPH, 26(3), 2007.
  62. [TBW+12] A. Tevs, A. Berner, M. Wand, I. Ihrke, M. Bokeloh, J. Kerber,
  63. Proceedings of ACM SIGGRAPH, 31(2), 2012.
  64. point set registration. In European Conference on Computer
  65. [Tsi03] Y. Tsin. Kernel Correlation as an Anity Measure in Point-
  66. Carnegie Mellon University, Pittsburgh, PA 15213, 2003.
  67. [WAO+09] M. Wand, B. Adams, M. Ovsjanikov, A. Berner, M. Bokeloh,
  68. scanner data. ACM Transactions on Graphics - Proceedings of
  69. ACM SIGGRAPH, 28(2), 2009.
  70. point clouds. In Proceedings of the fth Eurographics