Reconstruction of Polyhedron Objects by Structure Graph Integration - PowerPoint PPT Presentation

1 / 11
About This Presentation
Title:

Reconstruction of Polyhedron Objects by Structure Graph Integration

Description:

KOSYR'2001. Reconstruction of Polyhedron Objects. by Structure Graph Integration. Krzysztof Skabek ... to reconstruct unknown polyhedra even if viewpoints are ... – PowerPoint PPT presentation

Number of Views:69
Avg rating:3.0/5.0
Slides: 12
Provided by: Krzy1
Category:

less

Transcript and Presenter's Notes

Title: Reconstruction of Polyhedron Objects by Structure Graph Integration


1
Reconstruction of Polyhedron Objects by
Structure Graph Integration
  • Krzysztof Skabek
  • Institute of Theoretical and Applied Computer
    Sciences
  • Polish Academy of Sciences,
  • ul. Baltycka 5, 44-100 Gliwice,
  • e-mail krzysiek_at_iitis.gliwice.pl

2
Contents
  • 1. Stages of 3D Scene Reconstruction
  • 2. Graph Representation of 3D Scene
  • 3. Structure Graph Integration
  • 4. The Example of Scene Reconstruction
  • 5. Conclusions

3
Stages of 3D Scene Reconstruction
4
Graph Representation of 3D Scene
  • Contour Graph
  • Face Graph
  • set of vertices in the scene
  • set of connections among vertices
  • 3D coordinates (x,y,z) of vertices
  • set of faces in the scene
  • set of connections among faces
  • parameters of faces
  • parameters of face connections

5
Construction of Structure Graphs
4
7
4
Contour Graph
6
1
3
1
3
5
2
2
Face Graph
6
Integration of Structure Graphs I
  • GKi the contour graph obtained for ith view
    point
  • GSi the face graph for ith view created GKi
  • GKMi the integrated contour graph after i steps
  • GSM the integrated face graph after i steps
  • WKi set of corresponding vertices in GKi and
    GKMi-1
  • WSi set of corresponding vertices in GSi and
    GSMi-1

7
Integration of Structure Graphs II
  • Building GKi and GSi for ith view point
  • Subgraph matching of GSi and GSMi-1 storing the
    matched vertices in WSi.
  • Finding correspondences between GKi and WSi.
  • Calculation of the translation vector Ti and the
    rotation matrix Ri transformating coordinate
    system of GKi into the coordinates of GKMi-1
  • Calculation of new coordinates for vertices from
    GKi (using Ti and Ri)
  • Consolidation of contour graphsGKMi-1 GKi ?
    GKMi, GSMi-1 GSi ? GSMi
  • Checking the completeness of reconstruction

8
Completeness of the Representation
  • The total rotation angle in all iterations
    exceeds 360º
  • Passing a given number of iterations without
    encountering new scene objects

9
The Example of Scene Reconstruction
Input Graph
IntegratedContour Graph
10
Stages of Scene Reconstruction I
II.
III.
I.
11
Stages of Scene Reconstruction II
V.
VI.
IV.
12
Conclusions
  • Graph matching makes possible to reconstruct
    unknown polyhedra even if viewpoints are not
    exactly located in the scene
  • Heuristics significantly increase the average
    time of graph matching algorithms
  • The estimation of a viewpoint position is useful
    to verify the obtained scene models

13
The End
Write a Comment
User Comments (0)
About PowerShow.com