CLCC: Contrastive Learning for Color Constancy

Lo, Yi‐Chen; Chang, Chia-Che; Chiu, Hsuan-Chao; Huang, Yuhao; Chen, Chia-Ping; Chang, Youngho; Jou, Kevin

doi:10.1109/cvpr46437.2021.00796

Cited by 29 publications

(23 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Learning-based illuminant estimation methods include gamut-based methods [27,30,28,37], Bayesian methods [15,14,34,42], and neural network-based methods [31,20,55]. Advanced neural network-based methods further involve different learning strategies such as patch-wise learning [61,45], achromatic pixel detection [13], metric learning [67], contrastive learning [54], cross-camera illumination estimation [1] and weighting map blending [4].…”

Section: Illuminant Estimationmentioning

confidence: 99%

See 1 more Smart Citation

Modeling the Lighting in Scenes as Style for Auto White-Balance Correction

Kınlı¹,

Yılmaz²,

Özcan³

et al. 2022

Preprint

View full text Add to dashboard Cite

Style may refer to different concepts (e.g. painting style, hairstyle, texture, color, filter, etc.) depending on how the feature space is formed. In this work, we propose a novel idea of interpreting the lighting in the single-and multi-illuminant scenes as the concept of style. To verify this idea, we introduce an enhanced auto white-balance (AWB) method that models the lighting in single-and mixed-illuminant scenes as the style factor. Our AWB method does not require any illumination estimation step, yet contains a network learning to generate the weighting maps of the images with different WB settings. Proposed network utilizes the style information, extracted from the scene by a multi-head style extraction module. AWB correction is completed after blending these weighting maps and the scene. Experiments on single-and mixedilluminant datasets demonstrate that our proposed method achieves promising correction results when compared to the recent works. This shows that the lighting in the scenes with multiple illuminations can be modeled by the concept of style. Source code and trained models are available on https://github.com/birdortyedi/ lighting-as-style-awb-correction.

show abstract

Section: Illuminant Estimationmentioning

confidence: 99%

“…Prior works on AWB correction [61,45,10,12] thoroughly focuses on global illuminant estimation. The recent studies [67,42,54] achieve significant improvements on this task specialized on single-illuminant scenes. As a common practice, a diagonal-based correction matrix [38] is applied to the images to perform WB.…”

Section: Introductionmentioning

confidence: 99%

Modeling the Lighting in Scenes as Style for Auto White-Balance Correction

Kınlı¹,

Yılmaz²,

Özcan³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Contrastive Learning (CL) [11] helps learn robust feature spaces that are close across a data distribution and attributes that set apart a data distribution from another. CL has shown great promise in self-supervised regimes [6,9,13] while recently, it has also been applied to the supervised learning domain and achieved promising results [19,22,23]. CL has been used in a self-supervised manner to help debias models [21,26,32,37].…”

Section: Related Workmentioning

confidence: 99%

“…CL has been used in a self-supervised manner to help debias models [21,26,32,37]. In the fully supervised learning domain, previous works have shown that utilizing contrastive loss as an auxiliary loss can encourage learning more robust features with higher generalization abilities through careful contrastive pair construction [22,23]. To the best of our knowledge, we are the first to leverage contrastive learning as an auxiliary loss to improve the model's background robustness in a fully-supervised setting.…”

Section: Related Workmentioning

confidence: 99%

CLAD: A Contrastive Learning based Approach for Background Debiasing

Wang¹,

Machiraju²,

Choung³

et al. 2022

Preprint

View full text Add to dashboard Cite

Convolutional neural networks (CNNs) have achieved superhuman performance in multiple vision tasks, especially image classification. However, unlike humans, CNNs leverage spurious features, such as background information to make decisions. This tendency creates different problems in terms of robustness or weak generalization performance. Through our work, we introduce a contrastive learning-based approach (CLAD) to mitigate the background bias in CNNs. CLAD encourages semantic focus on object foregrounds and penalizes learning features from irrelavant backgrounds. Our method also introduces an efficient way of sampling negative samples. We achieve state-of-theart results on the Background Challenge dataset, outperforming the previous benchmark with a margin of 4.1%. Our paper shows how CLAD serves as a proof of concept for debiasing of spurious features, such as background and texture (in supplementary material).

show abstract

“…Computational color constancy tries to mimic this behaviour with computer vision systems. Given a digital image, the idea is first to estimate the color of the illumination and then to remove the impact of this illumination from the pixel colors [1,2,3,4]. In this paper, we propose to improve these two steps with an original and simple solution.…”

Section: Introductionmentioning

confidence: 99%

Predicting the Colors of Reference Surfaces for Color Constancy

Dubuisson

Muselet

Basso-Bert

et al. 2022

2022 IEEE International Conference on Image Processing (ICIP)

View full text Add to dashboard Cite

The classical color constancy algorithms concentrate only on the color of a grey surface to estimate the light color and to white balance the image. In this paper, we show that the quality of the whole process can be clearly improved by predicting and correcting the colors of a set of reference surfaces. The ground truth of the surface color under white light can be easily obtained with a set of images acquired by the considered camera under sun light. Thus, we design a deep network to predict the colors of the reference surfaces of a color checker as if it had been in the scene at acquisition time. We show that our solution improves the two steps of the color constancy process on 9 datasets and we claim that being able to synthetically insert a color chart in any image can help for many other tasks.

show abstract

CLCC: Contrastive Learning for Color Constancy

Cited by 29 publications

References 36 publications

Modeling the Lighting in Scenes as Style for Auto White-Balance Correction

Modeling the Lighting in Scenes as Style for Auto White-Balance Correction

CLAD: A Contrastive Learning based Approach for Background Debiasing

Predicting the Colors of Reference Surfaces for Color Constancy

Contact Info

Product

Resources

About