Saliency Map-Aided Generative Adversarial Network for RAW to RGB Mapping

Yuzhi Zhao; Lai-Man Po; Tiantian Zhang; Zongbang Liao; Xiang Shi; Yujia Zhang; Weifeng Ou; Pengfei Xian; Jingjing Xiong; Chang Zhou; Wing Yin Yu

doi:10.1109/ICCVW.2019.00428

Saliency Map-Aided Generative Adversarial Network for RAW to RGB Mapping

Yuzhi Zhao^*, Lai-Man Po, Tiantian Zhang, Zongbang Liao, Xiang Shi, Yujia Zhang, Weifeng Ou, Pengfei Xian, Jingjing Xiong, Chang Zhou, Wing Yin Yu

^*Corresponding author for this work

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

8 Citations (Scopus)

Abstract

RAW files are widely applied in cameras and scanners as storage because they contain original optical data. Different cameras usually process the RAW files using diverse algorithms that are incompatible. To address the issue, we propose a general transformation method for cross-camera RAW to RGB mapping based on Generative Adversarial Network (GAN). Moreover, we propose a saliency map-aided data augmentation technique and the saliency maps are produced by Saliency GAN (SalGAN). Given RAW file as an input, it jointly predicts the RGB image and corresponding saliency map to enhance perceptual quality in the generated image. The proposed architecture is trained on the Zurich RAW2RGB (ZRR) dataset. Experimental results show that our method can generate more clear and visually plausible images than state-of-the-art networks.

Original language	English
Title of host publication	Proceedings - 2019 International Conference on Computer Vision Workshops (ICCV 2019)
Publisher	IEEE
Pages	3449-3457
ISBN (Electronic)	978-1-7281-5023-9
DOIs	https://doi.org/10.1109/ICCVW.2019.00428
Publication status	Published - 27 Oct 2019
Event	17th IEEE/CVF International Conference on Computer Vision (ICCV 2019) - COEX Convention Center, Seoul, Korea, Republic of Duration: 27 Oct 2019 → 2 Nov 2019 http://iccv2019.thecvf.com/

Conference

Conference	17th IEEE/CVF International Conference on Computer Vision (ICCV 2019)
Abbreviated title	ICCV19
Country/Territory	Korea, Republic of
City	Seoul
Period	27/10/19 → 2/11/19
Internet address	http://iccv2019.thecvf.com/

Bibliographical note

Research Unit(s) information for this publication is provided by the author(s) concerned.

Research Keywords

Generative adversarial network
RAW to RGB
Saliency map

Access Output

10.1109/ICCVW.2019.00428

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@inproceedings{30be1eebbc8a4f51aaf0a7753b9e5612,

title = "Saliency Map-Aided Generative Adversarial Network for RAW to RGB Mapping",

abstract = "RAW files are widely applied in cameras and scanners as storage because they contain original optical data. Different cameras usually process the RAW files using diverse algorithms that are incompatible. To address the issue, we propose a general transformation method for cross-camera RAW to RGB mapping based on Generative Adversarial Network (GAN). Moreover, we propose a saliency map-aided data augmentation technique and the saliency maps are produced by Saliency GAN (SalGAN). Given RAW file as an input, it jointly predicts the RGB image and corresponding saliency map to enhance perceptual quality in the generated image. The proposed architecture is trained on the Zurich RAW2RGB (ZRR) dataset. Experimental results show that our method can generate more clear and visually plausible images than state-of-the-art networks.",

keywords = "Generative adversarial network, RAW to RGB, Saliency map",

author = "Yuzhi Zhao and Lai-Man Po and Tiantian Zhang and Zongbang Liao and Xiang Shi and Yujia Zhang and Weifeng Ou and Pengfei Xian and Jingjing Xiong and Chang Zhou and Yu, {Wing Yin}",

note = "Research Unit(s) information for this publication is provided by the author(s) concerned.; 17th IEEE/CVF International Conference on Computer Vision (ICCV 2019), ICCV19 ; Conference date: 27-10-2019 Through 02-11-2019",

year = "2019",

month = oct,

day = "27",

doi = "10.1109/ICCVW.2019.00428",

language = "English",

pages = "3449--3457",

booktitle = "Proceedings - 2019 International Conference on Computer Vision Workshops (ICCV 2019)",

publisher = "IEEE",

address = "United States",

url = "http://iccv2019.thecvf.com/",

}

Zhao, Y , Po, L-M , Zhang, T, Liao, Z, Shi, X, Zhang, Y , Ou, W , Xian, P , Xiong, J , Zhou, C & Yu, WY 2019, Saliency Map-Aided Generative Adversarial Network for RAW to RGB Mapping. in Proceedings - 2019 International Conference on Computer Vision Workshops (ICCV 2019)., 9022028, IEEE, pp. 3449-3457, 17th IEEE/CVF International Conference on Computer Vision (ICCV 2019), Seoul, Korea, Republic of, 27/10/19. https://doi.org/10.1109/ICCVW.2019.00428

Saliency Map-Aided Generative Adversarial Network for RAW to RGB Mapping. / Zhao, Yuzhi ; Po, Lai-Man ; Zhang, Tiantian et al.
Proceedings - 2019 International Conference on Computer Vision Workshops (ICCV 2019). IEEE, 2019. p. 3449-3457 9022028.

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - Saliency Map-Aided Generative Adversarial Network for RAW to RGB Mapping

AU - Zhao, Yuzhi

AU - Po, Lai-Man

AU - Zhang, Tiantian

AU - Liao, Zongbang

AU - Shi, Xiang

AU - Zhang, Yujia

AU - Ou, Weifeng

AU - Xian, Pengfei

AU - Xiong, Jingjing

AU - Zhou, Chang

AU - Yu, Wing Yin

N1 - Research Unit(s) information for this publication is provided by the author(s) concerned.

PY - 2019/10/27

Y1 - 2019/10/27

N2 - RAW files are widely applied in cameras and scanners as storage because they contain original optical data. Different cameras usually process the RAW files using diverse algorithms that are incompatible. To address the issue, we propose a general transformation method for cross-camera RAW to RGB mapping based on Generative Adversarial Network (GAN). Moreover, we propose a saliency map-aided data augmentation technique and the saliency maps are produced by Saliency GAN (SalGAN). Given RAW file as an input, it jointly predicts the RGB image and corresponding saliency map to enhance perceptual quality in the generated image. The proposed architecture is trained on the Zurich RAW2RGB (ZRR) dataset. Experimental results show that our method can generate more clear and visually plausible images than state-of-the-art networks.

AB - RAW files are widely applied in cameras and scanners as storage because they contain original optical data. Different cameras usually process the RAW files using diverse algorithms that are incompatible. To address the issue, we propose a general transformation method for cross-camera RAW to RGB mapping based on Generative Adversarial Network (GAN). Moreover, we propose a saliency map-aided data augmentation technique and the saliency maps are produced by Saliency GAN (SalGAN). Given RAW file as an input, it jointly predicts the RGB image and corresponding saliency map to enhance perceptual quality in the generated image. The proposed architecture is trained on the Zurich RAW2RGB (ZRR) dataset. Experimental results show that our method can generate more clear and visually plausible images than state-of-the-art networks.

KW - Generative adversarial network

KW - RAW to RGB

KW - Saliency map

UR - http://www.scopus.com/inward/record.url?scp=85082472181&partnerID=8YFLogxK

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-85082472181&origin=recordpage

U2 - 10.1109/ICCVW.2019.00428

DO - 10.1109/ICCVW.2019.00428

M3 - RGC 32 - Refereed conference paper (with host publication)

SP - 3449

EP - 3457

BT - Proceedings - 2019 International Conference on Computer Vision Workshops (ICCV 2019)

PB - IEEE

T2 - 17th IEEE/CVF International Conference on Computer Vision (ICCV 2019)

Y2 - 27 October 2019 through 2 November 2019

ER -

Saliency Map-Aided Generative Adversarial Network for RAW to RGB Mapping

Abstract

Conference

Bibliographical note

Research Keywords

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this