Nonuniform lattice regression for modeling the camera imaging pipeline

Hai Ting Lin, Zheng Lu, Seon Joo Kim, Michael S. Brown

Research output: Chapter in Book/Report/Conference proceedingConference contribution

7 Citations (Scopus)

Abstract

We describe a method to construct a sparse lookup table (LUT) that is effective in modeling the camera imaging pipeline that maps a RAW camera values to their sRGB output. This work builds on the recent in-camera color processing model proposed by Kim et al. [1] that included a 3D gamut-mapping function. The major drawback in [1] is the high computational cost of the 3D mapping function that uses radial basis functions (RBF) involving several thousand control points. We show how to construct a LUT using a novel nonuniform lattice regression method that adapts the LUT lattice to better fit the 3D gamut-mapping function. Our method offers not only a performance speedup of an order of magnitude faster than RBF, but also a compact mechanism to describe the imaging pipeline.

Original languageEnglish
Title of host publicationComputer Vision, ECCV 2012 - 12th European Conference on Computer Vision, Proceedings
Pages556-568
Number of pages13
EditionPART 1
DOIs
Publication statusPublished - 2012 Oct 30
Event12th European Conference on Computer Vision, ECCV 2012 - Florence, Italy
Duration: 2012 Oct 72012 Oct 13

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 1
Volume7572 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other12th European Conference on Computer Vision, ECCV 2012
CountryItaly
CityFlorence
Period12/10/712/10/13

Fingerprint

Look-up Table
Pipelines
Regression
Camera
Cameras
Imaging
Table lookup
Imaging techniques
Radial Functions
Modeling
Basis Functions
Control Points
Computational Cost
Speedup
Output
Color
Processing
Costs
Model

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Lin, H. T., Lu, Z., Kim, S. J., & Brown, M. S. (2012). Nonuniform lattice regression for modeling the camera imaging pipeline. In Computer Vision, ECCV 2012 - 12th European Conference on Computer Vision, Proceedings (PART 1 ed., pp. 556-568). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7572 LNCS, No. PART 1). https://doi.org/10.1007/978-3-642-33718-5_40
Lin, Hai Ting ; Lu, Zheng ; Kim, Seon Joo ; Brown, Michael S. / Nonuniform lattice regression for modeling the camera imaging pipeline. Computer Vision, ECCV 2012 - 12th European Conference on Computer Vision, Proceedings. PART 1. ed. 2012. pp. 556-568 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1).
@inproceedings{291aeb3d6b504148bc4f2ee54b93854e,
title = "Nonuniform lattice regression for modeling the camera imaging pipeline",
abstract = "We describe a method to construct a sparse lookup table (LUT) that is effective in modeling the camera imaging pipeline that maps a RAW camera values to their sRGB output. This work builds on the recent in-camera color processing model proposed by Kim et al. [1] that included a 3D gamut-mapping function. The major drawback in [1] is the high computational cost of the 3D mapping function that uses radial basis functions (RBF) involving several thousand control points. We show how to construct a LUT using a novel nonuniform lattice regression method that adapts the LUT lattice to better fit the 3D gamut-mapping function. Our method offers not only a performance speedup of an order of magnitude faster than RBF, but also a compact mechanism to describe the imaging pipeline.",
author = "Lin, {Hai Ting} and Zheng Lu and Kim, {Seon Joo} and Brown, {Michael S.}",
year = "2012",
month = "10",
day = "30",
doi = "10.1007/978-3-642-33718-5_40",
language = "English",
isbn = "9783642337178",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
number = "PART 1",
pages = "556--568",
booktitle = "Computer Vision, ECCV 2012 - 12th European Conference on Computer Vision, Proceedings",
edition = "PART 1",

}

Lin, HT, Lu, Z, Kim, SJ & Brown, MS 2012, Nonuniform lattice regression for modeling the camera imaging pipeline. in Computer Vision, ECCV 2012 - 12th European Conference on Computer Vision, Proceedings. PART 1 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 1, vol. 7572 LNCS, pp. 556-568, 12th European Conference on Computer Vision, ECCV 2012, Florence, Italy, 12/10/7. https://doi.org/10.1007/978-3-642-33718-5_40

Nonuniform lattice regression for modeling the camera imaging pipeline. / Lin, Hai Ting; Lu, Zheng; Kim, Seon Joo; Brown, Michael S.

Computer Vision, ECCV 2012 - 12th European Conference on Computer Vision, Proceedings. PART 1. ed. 2012. p. 556-568 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7572 LNCS, No. PART 1).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Nonuniform lattice regression for modeling the camera imaging pipeline

AU - Lin, Hai Ting

AU - Lu, Zheng

AU - Kim, Seon Joo

AU - Brown, Michael S.

PY - 2012/10/30

Y1 - 2012/10/30

N2 - We describe a method to construct a sparse lookup table (LUT) that is effective in modeling the camera imaging pipeline that maps a RAW camera values to their sRGB output. This work builds on the recent in-camera color processing model proposed by Kim et al. [1] that included a 3D gamut-mapping function. The major drawback in [1] is the high computational cost of the 3D mapping function that uses radial basis functions (RBF) involving several thousand control points. We show how to construct a LUT using a novel nonuniform lattice regression method that adapts the LUT lattice to better fit the 3D gamut-mapping function. Our method offers not only a performance speedup of an order of magnitude faster than RBF, but also a compact mechanism to describe the imaging pipeline.

AB - We describe a method to construct a sparse lookup table (LUT) that is effective in modeling the camera imaging pipeline that maps a RAW camera values to their sRGB output. This work builds on the recent in-camera color processing model proposed by Kim et al. [1] that included a 3D gamut-mapping function. The major drawback in [1] is the high computational cost of the 3D mapping function that uses radial basis functions (RBF) involving several thousand control points. We show how to construct a LUT using a novel nonuniform lattice regression method that adapts the LUT lattice to better fit the 3D gamut-mapping function. Our method offers not only a performance speedup of an order of magnitude faster than RBF, but also a compact mechanism to describe the imaging pipeline.

UR - http://www.scopus.com/inward/record.url?scp=84867853886&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84867853886&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-33718-5_40

DO - 10.1007/978-3-642-33718-5_40

M3 - Conference contribution

AN - SCOPUS:84867853886

SN - 9783642337178

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 556

EP - 568

BT - Computer Vision, ECCV 2012 - 12th European Conference on Computer Vision, Proceedings

ER -

Lin HT, Lu Z, Kim SJ, Brown MS. Nonuniform lattice regression for modeling the camera imaging pipeline. In Computer Vision, ECCV 2012 - 12th European Conference on Computer Vision, Proceedings. PART 1 ed. 2012. p. 556-568. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1). https://doi.org/10.1007/978-3-642-33718-5_40