18
Mesh Reduction with error control Scott Buffa Jeranfer Bermudez Alex Peer CSE 872 Implementation of (Klein, 2001)

Mesh Reduction with error control

  • Upload
    lorna

  • View
    66

  • Download
    0

Embed Size (px)

DESCRIPTION

Mesh Reduction with error control. Scott Buffa Jeranfer Bermudez Alex Peer CSE 872 Implementation of (Klein, 2001). Goals. Mesh simplification Hausdorff distance Constrained 3D Delaunay triangulation. Other solutions. Coplanar facets merging [HH92, MSS94] Mesh decimation: [SZL92] - PowerPoint PPT Presentation

Citation preview

Page 1: Mesh Reduction with error control

Mesh Reduction with error control

Scott BuffaJeranfer Bermudez

Alex PeerCSE 872

Implementation of (Klein, 2001)

Page 2: Mesh Reduction with error control

Goals Mesh simplification

› Hausdorff distance› Constrained 3D Delaunay triangulation

Page 3: Mesh Reduction with error control

Other solutions Coplanar facets merging [HH92, MSS94] Mesh decimation: [SZL92] Mesh optimization: [HDD93] Point coalescence: [RB93] Re-tiling: [Tur92] Multiresolution retiling: [EDD 95]

Drawback is no common way to measure the error between original and simplified meshes.

Page 4: Mesh Reduction with error control

Hausdorff distance Measures how far two subsets of space

are from each other› ›

› To get the maximum of the minimum distances betweentwo sets

Page 5: Mesh Reduction with error control

Delaunay Triangulation A triangulation such that no point is

inside a circumcircle of any triangle.› Maximizes the minimum angle of all angles

of triangles

Invented by Boris Delaunay, 1934

Page 6: Mesh Reduction with error control

Delaunay Triangulation cont.

Given three points:

› Radius:

› Circumsphere center : Where:

Page 7: Mesh Reduction with error control

Delaunay Triangulation cont.

Example:

› Points:

› Radius r = 3.605

› Center PC: = .5 = 0 = .5

017

,011

,051

321 PPPP1

P2 P3

y

x

Pc

017

5.011

0051

5.034

CP

Page 8: Mesh Reduction with error control

The Algorithm Overview

› Remove vertices and re-triangulate the resulting holes

› Stops when no vertices can be removed without exceeding the Hausdorff distance.

Page 9: Mesh Reduction with error control

The Algorithm cont. Iterative approach Determining the vertex to remove

› Hausdorff distance› Distance function› triangulation

Remove vertex› If we remove a vertex v from the

triangulation, its adjacent triangles are removed and the remaining hole is re-triangulated.

Page 10: Mesh Reduction with error control

The Algorithm cont. Hausdorff Distance

› For every vertex, simulate removal› Rank the vertices by the error that would

be introduced (hausdorff)› Remove vertex that is at top of list› Update error for vertices› Repeat removal and update until none can

be removed without exceeding maximum error tolerance

Page 11: Mesh Reduction with error control

The Algorithm cont. Delaunay Triangulation

› Find how vertex relates to triangle› If new vertex lies within triangle, split

triangle. Otherwise, connect to nearest points

› Use spheres to generate new faces

Page 12: Mesh Reduction with error control

What didn’t go well Constrained 3D Delaunay Triangulation

for face generation› Why did’t go well?

No examples. Poor documentation in 3D Needed ordered vertex list Overlapping faces

› Resorted to simple fan algorithm Generating list of vertices with proper

winding order

Page 13: Mesh Reduction with error control

Demo

Page 14: Mesh Reduction with error control

Results Simplified Mesh Original Mesh

Page 15: Mesh Reduction with error control

Results cont.

Page 16: Mesh Reduction with error control

Results cont.

Page 17: Mesh Reduction with error control

Results cont.

Page 18: Mesh Reduction with error control

Questions?