Approaching the computational color constancy as a classification problem through deep learning

Seoung Wug Oh, Seon Joo Kim

Research output: Contribution to journalArticle

26 Citations (Scopus)

Abstract

Computational color constancy refers to the problem of computing the illuminant color so that the images of a scene under varying illumination can be normalized to an image under the canonical illumination. In this paper, we adopt a deep learning framework for the illumination estimation problem. The proposed method works under the assumption of uniform illumination over the scene and aims for the accurate illuminant color computation. Specifically, we trained the convolutional neural network to solve the problem by casting the color constancy problem as an illumination classification problem. We designed the deep learning architecture so that the output of the network can be directly used for computing the color of the illumination. Experimental results show that our deep network is able to extract useful features for the illumination estimation and our method outperforms all previous color constancy methods on multiple test datasets.

Original languageEnglish
Pages (from-to)405-416
Number of pages12
JournalPattern Recognition
Volume61
DOIs
Publication statusPublished - 2017 Jan 1

Fingerprint

Lighting
Color
Deep learning
Casting
Neural networks

All Science Journal Classification (ASJC) codes

  • Software
  • Signal Processing
  • Computer Vision and Pattern Recognition
  • Artificial Intelligence

Cite this

@article{6c1e94e73fc6488798d32a6b5b1179ee,
title = "Approaching the computational color constancy as a classification problem through deep learning",
abstract = "Computational color constancy refers to the problem of computing the illuminant color so that the images of a scene under varying illumination can be normalized to an image under the canonical illumination. In this paper, we adopt a deep learning framework for the illumination estimation problem. The proposed method works under the assumption of uniform illumination over the scene and aims for the accurate illuminant color computation. Specifically, we trained the convolutional neural network to solve the problem by casting the color constancy problem as an illumination classification problem. We designed the deep learning architecture so that the output of the network can be directly used for computing the color of the illumination. Experimental results show that our deep network is able to extract useful features for the illumination estimation and our method outperforms all previous color constancy methods on multiple test datasets.",
author = "Oh, {Seoung Wug} and Kim, {Seon Joo}",
year = "2017",
month = "1",
day = "1",
doi = "10.1016/j.patcog.2016.08.013",
language = "English",
volume = "61",
pages = "405--416",
journal = "Pattern Recognition",
issn = "0031-3203",
publisher = "Elsevier Limited",

}

Approaching the computational color constancy as a classification problem through deep learning. / Oh, Seoung Wug; Kim, Seon Joo.

In: Pattern Recognition, Vol. 61, 01.01.2017, p. 405-416.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Approaching the computational color constancy as a classification problem through deep learning

AU - Oh, Seoung Wug

AU - Kim, Seon Joo

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Computational color constancy refers to the problem of computing the illuminant color so that the images of a scene under varying illumination can be normalized to an image under the canonical illumination. In this paper, we adopt a deep learning framework for the illumination estimation problem. The proposed method works under the assumption of uniform illumination over the scene and aims for the accurate illuminant color computation. Specifically, we trained the convolutional neural network to solve the problem by casting the color constancy problem as an illumination classification problem. We designed the deep learning architecture so that the output of the network can be directly used for computing the color of the illumination. Experimental results show that our deep network is able to extract useful features for the illumination estimation and our method outperforms all previous color constancy methods on multiple test datasets.

AB - Computational color constancy refers to the problem of computing the illuminant color so that the images of a scene under varying illumination can be normalized to an image under the canonical illumination. In this paper, we adopt a deep learning framework for the illumination estimation problem. The proposed method works under the assumption of uniform illumination over the scene and aims for the accurate illuminant color computation. Specifically, we trained the convolutional neural network to solve the problem by casting the color constancy problem as an illumination classification problem. We designed the deep learning architecture so that the output of the network can be directly used for computing the color of the illumination. Experimental results show that our deep network is able to extract useful features for the illumination estimation and our method outperforms all previous color constancy methods on multiple test datasets.

UR - http://www.scopus.com/inward/record.url?scp=84988484440&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84988484440&partnerID=8YFLogxK

U2 - 10.1016/j.patcog.2016.08.013

DO - 10.1016/j.patcog.2016.08.013

M3 - Article

VL - 61

SP - 405

EP - 416

JO - Pattern Recognition

JF - Pattern Recognition

SN - 0031-3203

ER -