bayesian probabilistic reasoning and learning part 2 huang jin state key lab of cad&cg zhejiang...

14
Bayesian Probabilistic reasoning and learning Part 2 Huang Jin State Key Lab of CAD&CG Zhejiang University

Upload: theodore-heath

Post on 17-Dec-2015

215 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Bayesian Probabilistic reasoning and learning Part 2 Huang Jin State Key Lab of CAD&CG Zhejiang University

Bayesian Probabilistic reasoning and learning

Part 2Huang Jin

State Key Lab of CAD&CGZhejiang University

Page 2: Bayesian Probabilistic reasoning and learning Part 2 Huang Jin State Key Lab of CAD&CG Zhejiang University

Problem

Page 3: Bayesian Probabilistic reasoning and learning Part 2 Huang Jin State Key Lab of CAD&CG Zhejiang University

Alpha Channel

• Instead of 0-1 mask.• What is F, B, and C?• What is alpha?• How to blend F, B into C by alpha?• How to get alpha from F, B, C?

• f: a projection, R3->R

Page 4: Bayesian Probabilistic reasoning and learning Part 2 Huang Jin State Key Lab of CAD&CG Zhejiang University

Known and Unknown

• Known: C• Unknown: F, B• MAP

– find the most likely estimates for F, B, and alpha, given the observation C.

• Why not

Page 5: Bayesian Probabilistic reasoning and learning Part 2 Huang Jin State Key Lab of CAD&CG Zhejiang University

Some Deduce

P(c) is constant

Page 6: Bayesian Probabilistic reasoning and learning Part 2 Huang Jin State Key Lab of CAD&CG Zhejiang University

L(C|F, B, alpha)

Giving F, B, alpha, what’s the probability of C?

Page 7: Bayesian Probabilistic reasoning and learning Part 2 Huang Jin State Key Lab of CAD&CG Zhejiang University

Gaussian Distribution

•Parameters: Mean and variance

•Multivariate Gaussian Distribution

P(x)=

Page 8: Bayesian Probabilistic reasoning and learning Part 2 Huang Jin State Key Lab of CAD&CG Zhejiang University

How to estimate F and B?

• From neighbor

• Weight the pixel’s contribution

• Multivariate: R, G, B

Page 9: Bayesian Probabilistic reasoning and learning Part 2 Huang Jin State Key Lab of CAD&CG Zhejiang University

L(F) and L(B)

distance related weight

Page 10: Bayesian Probabilistic reasoning and learning Part 2 Huang Jin State Key Lab of CAD&CG Zhejiang University

Put them togetherf(F, B, alpha)=

Page 11: Bayesian Probabilistic reasoning and learning Part 2 Huang Jin State Key Lab of CAD&CG Zhejiang University

Iteration

• Why not

• Not a quadratic equation.• we break the problem into two quadratic sub-

problems.• The second one:

Page 12: Bayesian Probabilistic reasoning and learning Part 2 Huang Jin State Key Lab of CAD&CG Zhejiang University

Additional Results

Page 13: Bayesian Probabilistic reasoning and learning Part 2 Huang Jin State Key Lab of CAD&CG Zhejiang University

Where is Bayesian?

• We need alpha:• We guess it under Gaussian Distribution :

Page 14: Bayesian Probabilistic reasoning and learning Part 2 Huang Jin State Key Lab of CAD&CG Zhejiang University

谢谢