Deep Dichromatic Model Estimation under AC Light Sources

Jun Sang Yoo; Chan Ho Lee; Jong Ok Kim

doi:10.1109/TIP.2021.3100550

Deep Dichromatic Model Estimation under AC Light Sources

Jun Sang Yoo, Chan Ho Lee, Jong Ok Kim

Research output: Contribution to journal › Article › peer-review

6 Citations (Scopus)

Abstract

The dichromatic reflection model has been popularly exploited for computer vison tasks, such as color constancy and highlight removal. However, dichromatic model estimation is an severely ill-posed problem. Thus, several assumptions have been commonly made to estimate the dichromatic model, such as white-light (highlight removal) and the existence of highlight regions (color constancy). In this paper, we propose a spatio-temporal deep network to estimate the dichromatic parameters under AC light sources. The minute illumination variations can be captured with high-speed camera. The proposed network is composed of two sub-network branches. From high-speed video frames, each branch generates chromaticity and coefficient matrices, which correspond to the dichromatic image model. These two separate branches are jointly learned by spatio-temporal regularization. As far as we know, this is the first work that aims to estimate all dichromatic parameters in computer vision. To validate the model estimation accuracy, it is applied to color constancy and highlight removal. Both experimental results show that the dichromatic model can be estimated accurately via the proposed deep network.

Original language	English
Article number	9508197
Pages (from-to)	7064-7073
Number of pages	10
Journal	IEEE Transactions on Image Processing
Volume	30
DOIs	https://doi.org/10.1109/TIP.2021.3100550
Publication status	Published - 2021
Externally published	Yes

Bibliographical note

Funding Information:
Manuscript received August 31, 2020; revised June 17, 2021; accepted July 13, 2021. Date of publication August 5, 2021; date of current version August 11, 2021. This work was supported in part by Samsung Electronics Company Ltd., under Grant Q1826052 and in part by the National Research Foundation of Korea (NRF) Grant funded by the Korean Government (MSIT) under Grant 2019R1A2C1005834. The associate editor coordinating the review of this manuscript and approving it for publication was Prof. Damon M. Chandler. (Corresponding author: Jong-Ok Kim.) Jun-Sang Yoo is with the Computer Vision Laboratory, Samsung Advanced Institute of Technology, Suwon, Gyeonggi-do 16678, South Korea (e-mail: [email protected]).

Publisher Copyright:
© 1992-2012 IEEE.

Keywords

AC light
Dichromatic reflection model
color constancy
highlight removal

ASJC Scopus subject areas

Software
Computer Graphics and Computer-Aided Design

Access to Document

10.1109/TIP.2021.3100550

Cite this

@article{63ea165b8e8f430b98ded77f5302cc13,

title = "Deep Dichromatic Model Estimation under AC Light Sources",

abstract = "The dichromatic reflection model has been popularly exploited for computer vison tasks, such as color constancy and highlight removal. However, dichromatic model estimation is an severely ill-posed problem. Thus, several assumptions have been commonly made to estimate the dichromatic model, such as white-light (highlight removal) and the existence of highlight regions (color constancy). In this paper, we propose a spatio-temporal deep network to estimate the dichromatic parameters under AC light sources. The minute illumination variations can be captured with high-speed camera. The proposed network is composed of two sub-network branches. From high-speed video frames, each branch generates chromaticity and coefficient matrices, which correspond to the dichromatic image model. These two separate branches are jointly learned by spatio-temporal regularization. As far as we know, this is the first work that aims to estimate all dichromatic parameters in computer vision. To validate the model estimation accuracy, it is applied to color constancy and highlight removal. Both experimental results show that the dichromatic model can be estimated accurately via the proposed deep network.",

keywords = "AC light, Dichromatic reflection model, color constancy, highlight removal",

author = "Yoo, {Jun Sang} and Lee, {Chan Ho} and Kim, {Jong Ok}",

note = "Funding Information: Manuscript received August 31, 2020; revised June 17, 2021; accepted July 13, 2021. Date of publication August 5, 2021; date of current version August 11, 2021. This work was supported in part by Samsung Electronics Company Ltd., under Grant Q1826052 and in part by the National Research Foundation of Korea (NRF) Grant funded by the Korean Government (MSIT) under Grant 2019R1A2C1005834. The associate editor coordinating the review of this manuscript and approving it for publication was Prof. Damon M. Chandler. (Corresponding author: Jong-Ok Kim.) Jun-Sang Yoo is with the Computer Vision Laboratory, Samsung Advanced Institute of Technology, Suwon, Gyeonggi-do 16678, South Korea (e-mail: [email protected]). Publisher Copyright: {\textcopyright} 1992-2012 IEEE.",

year = "2021",

doi = "10.1109/TIP.2021.3100550",

language = "English",

volume = "30",

pages = "7064--7073",

journal = "IEEE Transactions on Image Processing",

issn = "1057-7149",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - JOUR

T1 - Deep Dichromatic Model Estimation under AC Light Sources

AU - Yoo, Jun Sang

AU - Lee, Chan Ho

AU - Kim, Jong Ok

N1 - Funding Information: Manuscript received August 31, 2020; revised June 17, 2021; accepted July 13, 2021. Date of publication August 5, 2021; date of current version August 11, 2021. This work was supported in part by Samsung Electronics Company Ltd., under Grant Q1826052 and in part by the National Research Foundation of Korea (NRF) Grant funded by the Korean Government (MSIT) under Grant 2019R1A2C1005834. The associate editor coordinating the review of this manuscript and approving it for publication was Prof. Damon M. Chandler. (Corresponding author: Jong-Ok Kim.) Jun-Sang Yoo is with the Computer Vision Laboratory, Samsung Advanced Institute of Technology, Suwon, Gyeonggi-do 16678, South Korea (e-mail: [email protected]). Publisher Copyright: © 1992-2012 IEEE.

PY - 2021

Y1 - 2021

N2 - The dichromatic reflection model has been popularly exploited for computer vison tasks, such as color constancy and highlight removal. However, dichromatic model estimation is an severely ill-posed problem. Thus, several assumptions have been commonly made to estimate the dichromatic model, such as white-light (highlight removal) and the existence of highlight regions (color constancy). In this paper, we propose a spatio-temporal deep network to estimate the dichromatic parameters under AC light sources. The minute illumination variations can be captured with high-speed camera. The proposed network is composed of two sub-network branches. From high-speed video frames, each branch generates chromaticity and coefficient matrices, which correspond to the dichromatic image model. These two separate branches are jointly learned by spatio-temporal regularization. As far as we know, this is the first work that aims to estimate all dichromatic parameters in computer vision. To validate the model estimation accuracy, it is applied to color constancy and highlight removal. Both experimental results show that the dichromatic model can be estimated accurately via the proposed deep network.

AB - The dichromatic reflection model has been popularly exploited for computer vison tasks, such as color constancy and highlight removal. However, dichromatic model estimation is an severely ill-posed problem. Thus, several assumptions have been commonly made to estimate the dichromatic model, such as white-light (highlight removal) and the existence of highlight regions (color constancy). In this paper, we propose a spatio-temporal deep network to estimate the dichromatic parameters under AC light sources. The minute illumination variations can be captured with high-speed camera. The proposed network is composed of two sub-network branches. From high-speed video frames, each branch generates chromaticity and coefficient matrices, which correspond to the dichromatic image model. These two separate branches are jointly learned by spatio-temporal regularization. As far as we know, this is the first work that aims to estimate all dichromatic parameters in computer vision. To validate the model estimation accuracy, it is applied to color constancy and highlight removal. Both experimental results show that the dichromatic model can be estimated accurately via the proposed deep network.

KW - AC light

KW - Dichromatic reflection model

KW - color constancy

KW - highlight removal

UR - http://www.scopus.com/inward/record.url?scp=85112685346&partnerID=8YFLogxK

U2 - 10.1109/TIP.2021.3100550

DO - 10.1109/TIP.2021.3100550

M3 - Article

C2 - 34351857

AN - SCOPUS:85112685346

SN - 1057-7149

VL - 30

SP - 7064

EP - 7073

JO - IEEE Transactions on Image Processing

JF - IEEE Transactions on Image Processing

M1 - 9508197

ER -

Deep Dichromatic Model Estimation under AC Light Sources

Abstract

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this