Tampere University of Technology

TUTCRIS Research Portal

DR-GAN: Automatic Radial Distortion Rectification Using Conditional GAN in Real-Time

Research output: Contribution to journalArticleScientificpeer-review

Standard

DR-GAN: Automatic Radial Distortion Rectification Using Conditional GAN in Real-Time. / Liao, Kang; Lin, Chunyu; Zhao, Yao; Gabbouj, Moncef.

In: IEEE Transactions on Circuits and Systems for Video Technology, 07.02.2019.

Research output: Contribution to journalArticleScientificpeer-review

Harvard

Liao, K, Lin, C, Zhao, Y & Gabbouj, M 2019, 'DR-GAN: Automatic Radial Distortion Rectification Using Conditional GAN in Real-Time' IEEE Transactions on Circuits and Systems for Video Technology. https://doi.org/10.1109/TCSVT.2019.2897984

APA

Liao, K., Lin, C., Zhao, Y., & Gabbouj, M. (2019). DR-GAN: Automatic Radial Distortion Rectification Using Conditional GAN in Real-Time. IEEE Transactions on Circuits and Systems for Video Technology. https://doi.org/10.1109/TCSVT.2019.2897984

Vancouver

Liao K, Lin C, Zhao Y, Gabbouj M. DR-GAN: Automatic Radial Distortion Rectification Using Conditional GAN in Real-Time. IEEE Transactions on Circuits and Systems for Video Technology. 2019 Feb 7. https://doi.org/10.1109/TCSVT.2019.2897984

Author

Liao, Kang ; Lin, Chunyu ; Zhao, Yao ; Gabbouj, Moncef. / DR-GAN: Automatic Radial Distortion Rectification Using Conditional GAN in Real-Time. In: IEEE Transactions on Circuits and Systems for Video Technology. 2019.

Bibtex - Download

@article{d916b9289e0e4308a7f4e9428c931632,
title = "DR-GAN: Automatic Radial Distortion Rectification Using Conditional GAN in Real-Time",
abstract = "Radial distortion, which severely hinders object detection and semantic recognition, frequently exists in images captured using a wide-angle lens. Correction of this distortion of images is crucial in many computer vision applications. In this study, we present DR-GAN, a conditional generative adversarial network (GAN) for automatic radial distortion rectification (DR). To the best of our knowledge, this is the first end-to-end trainable adversarial framework for radial distortion rectification. DR-GAN trained using the proposed low-to-high perceptual loss learns the mapping relation between different structural images rather than estimating multifarious distortion parameters, while also realizing label-free training and one-stage rectification. As a benefit of one-stage rectification, the proposed method is extremely fast with the completion of rectification in real-time. This is approximately 22 × faster than the state-of-the-art methods. The experimental results show that DR-GAN achieves excellent performance in both quantitative measure (PSNR and SSIM) and visual qualitative appearance.",
author = "Kang Liao and Chunyu Lin and Yao Zhao and Moncef Gabbouj",
year = "2019",
month = "2",
day = "7",
doi = "10.1109/TCSVT.2019.2897984",
language = "English",
journal = "IEEE Transactions on Circuits and Systems for Video Technology",
issn = "1051-8215",
publisher = "Institute of Electrical and Electronics Engineers",

}

RIS (suitable for import to EndNote) - Download

TY - JOUR

T1 - DR-GAN: Automatic Radial Distortion Rectification Using Conditional GAN in Real-Time

AU - Liao, Kang

AU - Lin, Chunyu

AU - Zhao, Yao

AU - Gabbouj, Moncef

PY - 2019/2/7

Y1 - 2019/2/7

N2 - Radial distortion, which severely hinders object detection and semantic recognition, frequently exists in images captured using a wide-angle lens. Correction of this distortion of images is crucial in many computer vision applications. In this study, we present DR-GAN, a conditional generative adversarial network (GAN) for automatic radial distortion rectification (DR). To the best of our knowledge, this is the first end-to-end trainable adversarial framework for radial distortion rectification. DR-GAN trained using the proposed low-to-high perceptual loss learns the mapping relation between different structural images rather than estimating multifarious distortion parameters, while also realizing label-free training and one-stage rectification. As a benefit of one-stage rectification, the proposed method is extremely fast with the completion of rectification in real-time. This is approximately 22 × faster than the state-of-the-art methods. The experimental results show that DR-GAN achieves excellent performance in both quantitative measure (PSNR and SSIM) and visual qualitative appearance.

AB - Radial distortion, which severely hinders object detection and semantic recognition, frequently exists in images captured using a wide-angle lens. Correction of this distortion of images is crucial in many computer vision applications. In this study, we present DR-GAN, a conditional generative adversarial network (GAN) for automatic radial distortion rectification (DR). To the best of our knowledge, this is the first end-to-end trainable adversarial framework for radial distortion rectification. DR-GAN trained using the proposed low-to-high perceptual loss learns the mapping relation between different structural images rather than estimating multifarious distortion parameters, while also realizing label-free training and one-stage rectification. As a benefit of one-stage rectification, the proposed method is extremely fast with the completion of rectification in real-time. This is approximately 22 × faster than the state-of-the-art methods. The experimental results show that DR-GAN achieves excellent performance in both quantitative measure (PSNR and SSIM) and visual qualitative appearance.

U2 - 10.1109/TCSVT.2019.2897984

DO - 10.1109/TCSVT.2019.2897984

M3 - Article

JO - IEEE Transactions on Circuits and Systems for Video Technology

JF - IEEE Transactions on Circuits and Systems for Video Technology

SN - 1051-8215

ER -