en
0.25
0.5
0.75
1.25
1.5
1.75
2
Bayesian Nonparametric Intrinsic Image Decomposition
Published on Oct 29, 20143235 Views
We present a generative, probabilistic model that decomposes an image into reflectance and shading components. The proposed approach uses a Dirichlet process Gaussian mixture model where the mean para
Related categories
Chapter list
Bayesian NonparametricIntrinsic Image Decomposition00:00
Shading & Reflectance00:06
Previous Approaches00:40
Previous Approach: SIRFS01:18
Previous Approach: Retinex - 102:22
Previous Approach: Retinex - 202:45
Previous Approach: Retinex - 302:46
Previous Approach: Retinex - 402:48
Previous Approach: Retinex - 503:02
Previous Approach: Retinex - 603:07
Advancement: Reflectance Sparsity03:23
Previous Approach: Gehler et al. NIPS’11 - 103:50
Previous Approach: Gehler et al. NIPS’11 - 204:42
Our Idea - 106:27
Our Idea - 206:34
Our Idea - 306:40
Our Idea - 406:45
Our Idea - 507:07
Our Idea - 607:49
Our Idea - 708:05
Our Idea - 808:25
Our Idea - 909:39
Our Idea - 1010:05
Our Idea - 1110:11
Our Idea - 1211:27
Our Idea - 1312:18
Improved Results13:25
Failure Cases14:09
Results - 114:43
Results - 214:49
Summary15:24