Download presentation
Presentation is loading. Please wait.
Published byWilliam Owens Modified over 8 years ago
1
Dynamic Geometry Registration Maks Ovsjanikov Leonidas Guibas Natasha GelfandNiloy J. Mitra Simon Flöry Helmut Pottmann Dynamic Geometry Registration
2
Scan Registration
3
Dynamic Geometry Registration Scan Registration Solve for inter-frame motion:
4
Dynamic Geometry Registration Scan Registration Solve for inter-frame motion:
5
Dynamic Geometry Registration The Setup Given: A set of frames {P 0, P 1,... P n } Goal: Recover rigid motion { 1, 2,... n } between adjacent frames
6
Dynamic Geometry Registration The Setup Smoothly varying object motion Unknown correspondence between scans Fast acquisition ! motion happens between frames
7
Dynamic Geometry Registration Contributions Rigid registration ! kinematic property of space- time surface (locally exact) Registration ! surface normal estimation Analysis (see paper for details) Relation to ICP (Iterated Closest Point) Stability analysis Extension to deformable/articulated bodies
8
Dynamic Geometry Registration Time Ordered Scans tjtj t j+1 t j+2
9
Dynamic Geometry Registration Space-time Surface
10
Dynamic Geometry Registration Space-time Surface !
11
Dynamic Geometry Registration Space-time Surface !
12
Dynamic Geometry Registration Spacetime Velocity Vectors Tangential point movement ! velocity vectors orthogonal to surface normals
13
Dynamic Geometry Registration Spacetime Velocity Vectors Tangential point movement ! velocity vectors orthogonal to surface normals
14
Dynamic Geometry Registration Final Steps (rigid) velocity vectors !
15
Dynamic Geometry Registration Final Steps (rigid) velocity vectors !
16
Dynamic Geometry Registration Registration Algorithm 1.Compute time coordinate spacing ( ), and form space-time surface. 2.Compute space time neighborhood using ANN, and locally estimate space-time surface normals. 3.Solve linear system to estimate (c j,c j ). 4.Convert velocity vectors to rotation matrix + translation vector using Plücker coordinates and quarternions.
17
Dynamic Geometry Registration Normal Estimation: PCA Based Plane fitting using PCA using chosen neighborhood points.
18
Dynamic Geometry Registration Normal Estimation: Iterative Refinement Update neighborhood with current velocity estimate.
19
Dynamic Geometry Registration Normal Refinement: Effect of Noise Stable, but more expensive.
20
Dynamic Geometry Registration Normal Estimation: Local Triangulation Perform local surface triangulation (tetrahedralization).
21
Dynamic Geometry Registration Normal Estimation Stable, but more expensive.
22
Dynamic Geometry Registration Comparison with ICP ICP point-planeDynamic registration
23
Dynamic Geometry Registration Rigid: Bee Sequence (2,200 frames)
24
Dynamic Geometry Registration Rigid: Coati Sequence (2,200 frames)
25
Dynamic Geometry Registration Handling Large Number of Frames
26
Dynamic Geometry Registration Rigid/Deformable: Teapot Sequence (2,200 frames)
27
Dynamic Geometry Registration Deformable Bodies Cluster points, and solve smaller systems. Propagate solutions with regularization.
28
Dynamic Geometry Registration Deformable: Hand (100 frames)
29
Dynamic Geometry Registration Deformable: Hand (100 frames) scan #1 ! scan #50 scan #1 ! scan #100
30
Dynamic Geometry Registration Deformation + scanner motion: Skeleton (100 frames)
31
Dynamic Geometry Registration Deformation + scanner motion: Skeleton (100 frames) scan #1 ! scan #50 scan #1 ! scan #100
32
Dynamic Geometry Registration Deformation + scanner motion: Skeleton (100 frames) rigid components
33
Dynamic Geometry Registration Performance Table (on 2.4GHz Athlon Dual Core, 2GB RAM)
34
Dynamic Geometry Registration Conclusion Simple algorithm using kinematic properties of space-time surface. Easy modification for deformable bodies. Suitable for use with fast scanners.
35
Dynamic Geometry Registration Conclusion Simple algorithm using kinematic properties of space-time surface. Easy modification for deformable bodies. Suitable for use with fast scanners. Limitations/Future Work Need more scans, dense scans, … Sampling condition ! time and space
36
Dynamic Geometry Registration Acknowledgements Thibaut Weise Andreas Schilling Funding sources: Austrian Science Fund grants S9206-N12 and P18865-N13 Neumaier fellowship NSF ITR 0205671 NIH grant GM-072970 DARPA grant 32905 Agilent Corporation
37
Dynamic Geometry Registration thank you
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.