3d descriptor design and learning for robust non rigid
play

3D Descriptor Design and Learning for Robust Non-rigid Shape - PowerPoint PPT Presentation

3D Descriptor Design and Learning for Robust Non-rigid Shape Matching Jianwei Guo NLPR, Institute of Automation, Chinese Academy of Sciences Sept.17, 2020 GAMES Webinar 1 Team--3D Visual Computing [ACM TOG (SIGGRAPH Asia) 2012]


  1. 3D Descriptor Design and Learning for Robust Non-rigid Shape Matching Jianwei Guo 郭建伟 NLPR, Institute of Automation, Chinese Academy of Sciences Sept.17, 2020 GAMES Webinar 1

  2. Team--3D Visual Computing [ACM TOG (SIGGRAPH Asia) 2012] [GMP 2006, CAD 2012, JCAD 2018] [ACM TOG (SigAsia) 2016] [CAGD 2016] [TVCG 2017] [ CAGD 2018 ] [ IEEE TIFS 2018 ] [CAD 2020] [CGF 2018] [CAD 2019] [TVCG 2018] [CGF 2019] [TVCG 2020] [TVCG 2020] [ACM TOG 2020] [ECCV 2018] [CVPR 2019] [ ACM TOG (SIGGRAPH) 2020] Mode delin ing & Mesh sh Optimi imizatio zation Shape Analysis alysis 2

  3. 3D Shape Descriptor Local descriptor Global descriptor 3

  4. 3D Shape Descriptor Descriptor design goals: Discriminative: able to determine if a pair • of vertices is similar or different Robust: work with different discretizations • of a surface Local descriptor 4

  5. Applications Shape matching/correspondence Shape retrieval [Ovsjanikov et al. 2017;Wang et al. 2018] [Ovsjanikov et al. 2009] Shape segmentation Surface registration [Rustamov 2007] 5 [Lui et al. 2010]

  6. Shape matching/correspondence 6

  7. Shape matching/correspondence Van Kaick et al.: A survey on shape correspondence . CGF, 2011. 7 Ovsjanikov et al.: Computing and processing correspondences with functional maps . SIGGRAPH ASIA 2016 Courses.

  8. Related Work Spatial domain approaches ❖ SI [Johnson and Hebert 3DSC [Frome et al. 2004] RoPS [Guo et al. 2013] SHOT [Tombari et al. 2010] 1999] PFH [Rusu et al. 2008] TriSI [Guo et al. 2015] FPFH [Rusu et al. 2009] Mesh-HOG [Zaharescu et al. 2009] Guo et al.: A comprehensive performance evaluation of 3D local feature descriptors . IJCV, 2016. 8

  9. Related Work Spectral domain approaches ❖ Shape-DNA [1999Reuter et al. GPS [Rustamov 2007 ] HKS [Sun et al. 2010] 2006] Scale-invariant HKS [Bronstein WKS [Aubry et al. 2011] DTEP [Melziet al. 2018] and Kokkinos 2010] 9

  10. Related Work Deep learning approaches ❖ [Zeng et al. 2017] [Huang et al. 2018] CGF [Khoury et al. 2017] PPFNet [Deng et al. 2018] • Mesh-aware OSD [Litman and Bronstein 2014] [Boscaini et al. 2015, 2016] [Masci et al. 2018] 10

  11. Related Work Non-learned descriptors ❖ Supervised descriptors ❖ • + pros : : WKS and HKS are robust • + pros: GNNs compute discriminative descriptors • - cons : not as discriminative as • - cons : less robust to different surface supervised descriptors discretizations 11

  12. Datasets SCAPE [Anguelov et al. 2005] TOSCA [Bronstein et al. 2008] SPRING [Yang et al. 2014] FAUST [Bogo et al. 2014] 12

  13. Our Work Contributions ❖ — Two new non-learned features, namely, Local Point Signature (LPS) and Wavelet Energy Decomposition Signature (WEDS). They exhibit high resilience to changes in mesh resolution, triangulation, scale, and rotation. — Two supervised frameworks to transform the non-learned features to more discriminative descriptors [ECCV 2018] [CVPR 2019] [ CVMJ 2020 ] [ ACM TOG (SIGGRAPH) 2020] 13

  14. 1-Descriptor Learning using Geometry Images Motivation ❖ Geometry images [Gu et al. 2002] 14

  15. 1-Descriptor Learning using Geometry Images Motivation ❖ Desbrun et al.: Intrinsic parameterizations of surface meshes . CGF, 2002 15

  16. 1-Descriptor Learning using Geometry Images Triplet Neural Network ❖ 16

  17. 1-Descriptor Learning using Geometry Images Triplet Loss ❖ • • Classic triplet loss: Min-Coefficient of Variation (Min-CV) loss: 17

  18. 1-Descriptor Learning using Geometry Images Results: shape matching ❖ 18

  19. 1-Descriptor Learning using Geometry Images Discrete Geodesic Polar Coordinates (DGPC) ❖ ill-shaped triangles degenerate triangles DGPC [ Melvær and Reimers 2012 ] Melvær and Reimers. Geodesic polar coordinates on polygonal meshes . CGF, 2012 19

  20. 1-Descriptor Learning using Geometry Images Results : shape correspondence ❖ • Cumulative match characteristic (CMC) • Cumulative geodesic error ( CGE ) FAUST SPRING 20

  21. 1-Descriptor Learning using Geometry Images Results : shape correspondence ❖ SCAPE 21

  22. 1-Descriptor Learning using Geometry Images Results : shape correspondence ❖ FAUST 22

  23. 2-Robust Local Spectral Descriptor Motivation ❖ Approach? Dataset? Resolutions, triangulations, scales, rotations 23

  24. 2-Robust Local Spectral Descriptor Motivation ❖ Dirichlet energy → : S f R ( ) ( ) ( ) ( )   2 =  =  E f f v dv f v f v dv 24 S S

  25. 2-Robust Local Spectral Descriptor Local Point Signature ❖ ( ) ( ) ( ) ( )   2 =  =  ( ) E f f v dv f v f v dv  = −  div f f S S Laplace – Beltrami operator  +   − cot cot ij ij  if , i j areadjacent 2 a  i   +  cot cot  = =  ik ik L if i j ij 2  a k i  0 otherwise    +  ( ) cot cot ( ) 1  ( ) ( )  =   ij ij  =   = − -  = − Mesh f v f v f v L , 0,1,..., 1 | 0,1,..., 1 M i i k i i k i i j i i 2 a ( )  → j N v i : f V R i 25

  26. 2-Robust Local Spectral Descriptor Local Point Signature ❖  =   = −   =   , 0,1,..., 1 T T A i i k , A i i i j i j A Spectral coefficients − ( ) N 1 =  =    =  =  T 2 AL T E f f f , f f A j j j j j A j =0 − − ( ) 1 1 d N N d    = → =   =   d 2 2 ( , ,..., ) : E F F f f f V R ij j j ij 1 2 d = = i 1 j =0 j =0 i 1   d d d    =        2 2 2 , ,..., sf  − − 1 1 2 2 1 1 i i N iN   = = = i 1 i 1 i 1 26

  27. 2-Robust Local Spectral Descriptor Local Point Signature ❖ • Build a local patch mesh around a vertex • Compute Laplacian eigenvectors and eigenvalues • Compute spectral coefficients  =  =  = → T , 3 f f A ( , , ): X x x x V R j j j 1 2 3 A • LPS is derived from Dirichlet energy and expressed as follows   3 3 3    =         2 2 2 , ,.., LPS   1 1 2 2 16 16 i i i   = = = 1 1 1 i i i 27

  28. 2-Robust Local Spectral Descriptor Descriptor learning ❖ 28

  29. 2-Robust Local Spectral Descriptor Descriptor learning ❖ 29

  30. 2-Robust Local Spectral Descriptor Dataset ❖ Discrete optimization method ( Wang et al. 2019) 30

  31. 2-Robust Local Spectral Descriptor Results: Robust to resolution and triangulation ❖ 31

  32. 2-Robust Local Spectral Descriptor Results: Comparison ❖ 32

  33. 3-Descriptor Learning using Multiscale GCNs Motivation ❖ – Time-consuming – Miss global information cut a geodesic disk 34

  34. 3-Descriptor Learning using Multiscale GCNs Graph wavelets ❖ Wavelet filter functions Scaling functions Wavelet functions • Multiscale property of wavelet functions 35

  35. 3-Descriptor Learning using Multiscale GCNs Graph wavelets ❖ • Reconstruction property of multiscale wavelet functions • Robustness to the change of resolution and triangulation 36

  36. 3-Descriptor Learning using Multiscale GCNs Wavelet Energy Decomposition Signature (WEDS) ❖ Reconstruct discrete Dirichlet energy Restructure into a sum per vertex Collect the local energy using multiscale wavelets 37

  37. 3-Descriptor Learning using Multiscale GCNs Wavelet Energy Decomposition Signature (WEDS) ❖ 38

  38. 3-Descriptor Learning using Multiscale GCNs Multiscale Graph Convolution Network ❖ • Convolutions on graphs are defined as: • ChebyNet approximates using 𝑛 -order polynomials: • MGCN approximates using multiscale wavelet filter basis: 39

  39. 3-Descriptor Learning using Multiscale GCNs Multiscale Graph Convolution Network ❖ • The multiscale convolution can be simplified as follows: Robustness to discretization Convolve local and global information 40

  40. 3-Descriptor Learning using Multiscale GCNs Results: performance of WEDS ❖ 41

  41. 3-Descriptor Learning using Multiscale GCNs Results: shape correspondence ❖ FAUST 42

  42. 3-Descriptor Learning using Multiscale GCNs Results: robust to resolution and triangulation ❖ 43

  43. 3-Descriptor Learning using Multiscale GCNs Results: robust to resolution and triangulation ❖ Extended FAUST 44

  44. 3-Descriptor Learning using Multiscale GCNs Results: non-isometric shapes ❖ 45

  45. Conclusions Contributions ❖ — Two non-learned features: Local Point Signature (LPS) and Wavelet Energy Decomposition Signature (WEDS). — Two supervised frameworks to transform the non-learned features to more discriminative descriptors Future work ❖ — Extend to point clouds and triangle soups — Industrial applications 46

  46. Acknowledgements Co-authors: CCF- 腾讯犀牛鸟科研基金 Dong-Ming Yan , CASIA Yiqun Wang , CASIA Hanyu Wang , University of Maryland-College Park Jing Ren, Peter Wonka , KAUST Code and Data: LDGI: https://github.com/jianweiguo/local3Ddescriptorlearning LPS: https://github.com/yiqun-wang/LPS MGCN: https://github.com/yiqun-wang/MGCN Thank you! 47

Recommend


More recommend