andrew nealen tu berlin takeo igarashi the university of tokyo / presto jst olga sorkine

Post on 19-Jan-2016

28 Views

Category:

Documents

1 Downloads

Preview:

Click to see full reader

DESCRIPTION

Laplacian Mesh Optimization. Andrew Nealen TU Berlin Takeo Igarashi The University of Tokyo / PRESTO JST Olga Sorkine Marc Alexa TU Berlin. What is it ?. Overview. Motivation Problem formulation Laplacian mesh processing basics Laplacian mesh optimization framework Applications - PowerPoint PPT Presentation

TRANSCRIPT

Andrew Nealen, TU Berlin, 2006 1

CG

11

Andrew NealenTU Berlin

Takeo IgarashiThe University of Tokyo / PRESTO JST

Olga SorkineMarc Alexa

TU Berlin

Laplacian Mesh Optimization

Laplacian Mesh Optimization

Andrew Nealen, TU Berlin, 2006 2

CG

22

What is it ?

Andrew Nealen, TU Berlin, 2006 3

CG

33

Overview

Motivation• Problem formulation• Laplacian mesh processing basics

Laplacian mesh optimization framework Applications

• Triangle shape optimization• Mesh smoothing

Discussion

Andrew Nealen, TU Berlin, 2006 4

CG

44

Motivation

Local detail preserving triangle optimization• A Sketch-Based Interface for Detail Preserving

Mesh Editing [Nealen et al. 2005]

Andrew Nealen, TU Berlin, 2006 5

CG

55

Motivation

Local detail preserving triangle optimization• A Sketch-Based Interface for Detail Preserving

Mesh Editing [Nealen et al. 2005]

Can we perform global optimization this way ?

=L x

Andrew Nealen, TU Berlin, 2006 6

CG

66

Laplacian Mesh Processing

Discrete Laplacians

=L x

n

cotangent : wij = cot ij + cot ij

uniform : wij = 1

( , )( , )

1i i ij j

i j Eiji j E

ww

δ x x

Andrew Nealen, TU Berlin, 2006 7

CG

77

Laplacian Mesh Processing

Surface reconstructionn

cotangent : wij = cot ij + cot ij

( , )( , )

1i i ij j

i j Eiji j E

ww

δ x x

uniform : wij = 1

=L x

L

L

y

z

x

z

y

x

Andrew Nealen, TU Berlin, 2006 8

CG

88

Laplacian Mesh Processing

Surface reconstructionn

z

y

x

y

z

x

=L

L

L

c1

fixedit

c2

Andrew Nealen, TU Berlin, 2006 9

CG

99

Laplacian Mesh Processing

Least-squares solutionn

z

y

x

y

z

x

=L

L

L

c1

fixedit

c2

w1 w1

w2 w2

wLi wLi

A x = bATA x = bAT

(ATA)-1x = bAT

Normal Equations

Andrew Nealen, TU Berlin, 2006 10

CG

1010

Laplacian Mesh Processing

Tangential smoothingn

z

y

x

y

z

x

=L

L

L

fixc1

L

L

L

Andrew Nealen, TU Berlin, 2006 11

CG

1111

L

L

L

Laplacian Mesh Processing

Tangential smoothingn

z

y

x

y

z

x

=

fixc1

Andrew Nealen, TU Berlin, 2006 12

CG

1212

L

L

L

Laplacian Mesh Processing

Tangential smoothingn

z

y

x

y

z

x

=

fixc1

Andrew Nealen, TU Berlin, 2006 13

CG

1313

More motivation…

So: can we use such a system for globaloptimization ?

=L x

Andrew Nealen, TU Berlin, 2006 14

CG

1414

Our Solution

All vertices are (weighted) anchors

Preserves global shape Uses existing LS framework Anchor + Laplacian weights determine

result

Andrew Nealen, TU Berlin, 2006 15

CG

1515

Framework

Detail preserving tri shape optimization for L = Luni and f = cot(similar to local optimization)

Mesh smoothing L = Lcot (outer fairness) or L = Luni (outer and inner fairness) and f = 0

=L x fWL WL

pWP WP

Andrew Nealen, TU Berlin, 2006 16

CG

1616

Tri Shape Optimization

Detail preserving tri shape optimization

=Luni x

pWP WP

Andrew Nealen, TU Berlin, 2006 17

CG

1717

Positional Weights

Andrew Nealen, TU Berlin, 2006 18

CG

1818

Constant Weights

Andrew Nealen, TU Berlin, 2006 19

CG

1919

Linear Weights

Andrew Nealen, TU Berlin, 2006 20

CG

2020

CDF Weights

Andrew Nealen, TU Berlin, 2006 21

CG

2121

CDF Weights

Andrew Nealen, TU Berlin, 2006 22

CG

2222

Sharp Features

Andrew Nealen, TU Berlin, 2006 23

CG

2323

Sharp Features

Andrew Nealen, TU Berlin, 2006 24

CG

2424

Sharp Features

Andrew Nealen, TU Berlin, 2006 25

CG

2525

Mesh Smoothing

Mesh smoothing L = Lcot (outer fairness) or L = Lumb (outer and inner fairness) and f = 0

Controlled by WP and WL (Intensity, Features) Similar to Least-Squares Meshes [Sorkine et al. 04]

=L x 0WL WL

pWP WP

Andrew Nealen, TU Berlin, 2006 26

CG

2626

Using WP

Andrew Nealen, TU Berlin, 2006 27

CG

2727

Using WP and WL

Andrew Nealen, TU Berlin, 2006 28

CG

2828

Results

Andrew Nealen, TU Berlin, 2006 29

CG

2929

Noisy

Andrew Nealen, TU Berlin, 2006 30

CG

3030

Smoothed

Andrew Nealen, TU Berlin, 2006 31

CG

3131

Original

Andrew Nealen, TU Berlin, 2006 32

CG

3232

Tri Shape Optimization

Andrew Nealen, TU Berlin, 2006 33

CG

3333

Smoothing Outer and Inner Fairness (Lumb)

Andrew Nealen, TU Berlin, 2006 34

CG

3434

Original

Andrew Nealen, TU Berlin, 2006 35

CG

3535

Tri Shape Optimization

Andrew Nealen, TU Berlin, 2006 36

CG

3636

SmoothingOuter Fairness only (Lcot)

Andrew Nealen, TU Berlin, 2006 37

CG

3737

Discussion

The good...• Easily controllable tri shape optimization and

smoothing• Leverages existing least squares framework• Can replace tangential smoothing step for

general remeshers ... and the not so good

• Euclidean distance is not Hausdorff distance, so error control is indirect

• Does rely on some (user) parameter tweaking

Andrew Nealen, TU Berlin, 2006 38

CG

3838

Thank you !

Contact info

Andrew Nealennealen@cs.tu-berlin.de

Takeo Igarashitakeo@acm.org

Olga Sorkinesorkine@cs.tu-berlin.de

Marc Alexamarc@cs.tu-berlin.de

top related