Abstract
We propose a new technique for estimating spatially varying parametric materials from a single image of an object with unknown shape in unknown illumination. Our method uses a low-order parametric reflectance model, and incorporates strong assumptions about lighting and shape. We develop new priors about how materials mix over space, and jointly infer all of these properties from a single image. This produces a decomposition of an image which corresponds, in one sense, to microscopic features (material reflectance) and macroscopic features (weights defining the mixing properties of materials over space). We have built a large dataset of real objects rendered with different material models under different illumination fields for training and ground truth evaluation. Extensive experiments on both our synthetic dataset images as well as real images show that (a) our method recovers parameters with reasonable accuracy; (b) material parameters recovered by our method give accurate predictions of new renderings of the object; and (c) our low-order reflectance model still provides a good fit to many real-world reflectances.
| Original language | English (US) |
|---|---|
| Title of host publication | SIGGRAPH Asia 2014 Indoor Scene Understanding Where Graphics Meets Vision, SA 2014 |
| Publisher | Association for Computing Machinery, Inc |
| ISBN (Electronic) | 9781450332422 |
| DOIs | |
| State | Published - Nov 24 2014 |
| Event | SIGGRAPH Asia 2014 Workshop on Indoor Scene Understanding Where Graphics Meets Vision, SA 2014 - Shenzhen, China Duration: Dec 3 2014 → Dec 6 2014 |
Publication series
| Name | SIGGRAPH Asia 2014 Indoor Scene Understanding Where Graphics Meets Vision, SA 2014 |
|---|
Other
| Other | SIGGRAPH Asia 2014 Workshop on Indoor Scene Understanding Where Graphics Meets Vision, SA 2014 |
|---|---|
| Country | China |
| City | Shenzhen |
| Period | 12/3/14 → 12/6/14 |
Fingerprint
Keywords
- Material modeling
- Material transfer
- Reflectance estimation
- Shape from shading
ASJC Scopus subject areas
- Computer Graphics and Computer-Aided Design
- Software
Cite this
Blind recovery of spatially varying reflectance from a single image. / Karsch, Kevin; Forsyth, David.
SIGGRAPH Asia 2014 Indoor Scene Understanding Where Graphics Meets Vision, SA 2014. Association for Computing Machinery, Inc, 2014. 2 (SIGGRAPH Asia 2014 Indoor Scene Understanding Where Graphics Meets Vision, SA 2014).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Blind recovery of spatially varying reflectance from a single image
AU - Karsch, Kevin
AU - Forsyth, David
PY - 2014/11/24
Y1 - 2014/11/24
N2 - We propose a new technique for estimating spatially varying parametric materials from a single image of an object with unknown shape in unknown illumination. Our method uses a low-order parametric reflectance model, and incorporates strong assumptions about lighting and shape. We develop new priors about how materials mix over space, and jointly infer all of these properties from a single image. This produces a decomposition of an image which corresponds, in one sense, to microscopic features (material reflectance) and macroscopic features (weights defining the mixing properties of materials over space). We have built a large dataset of real objects rendered with different material models under different illumination fields for training and ground truth evaluation. Extensive experiments on both our synthetic dataset images as well as real images show that (a) our method recovers parameters with reasonable accuracy; (b) material parameters recovered by our method give accurate predictions of new renderings of the object; and (c) our low-order reflectance model still provides a good fit to many real-world reflectances.
AB - We propose a new technique for estimating spatially varying parametric materials from a single image of an object with unknown shape in unknown illumination. Our method uses a low-order parametric reflectance model, and incorporates strong assumptions about lighting and shape. We develop new priors about how materials mix over space, and jointly infer all of these properties from a single image. This produces a decomposition of an image which corresponds, in one sense, to microscopic features (material reflectance) and macroscopic features (weights defining the mixing properties of materials over space). We have built a large dataset of real objects rendered with different material models under different illumination fields for training and ground truth evaluation. Extensive experiments on both our synthetic dataset images as well as real images show that (a) our method recovers parameters with reasonable accuracy; (b) material parameters recovered by our method give accurate predictions of new renderings of the object; and (c) our low-order reflectance model still provides a good fit to many real-world reflectances.
KW - Material modeling
KW - Material transfer
KW - Reflectance estimation
KW - Shape from shading
UR - http://www.scopus.com/inward/record.url?scp=84919389347&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84919389347&partnerID=8YFLogxK
U2 - 10.1145/2670291.2670293
DO - 10.1145/2670291.2670293
M3 - Conference contribution
AN - SCOPUS:84919389347
T3 - SIGGRAPH Asia 2014 Indoor Scene Understanding Where Graphics Meets Vision, SA 2014
BT - SIGGRAPH Asia 2014 Indoor Scene Understanding Where Graphics Meets Vision, SA 2014
PB - Association for Computing Machinery, Inc
ER -