Convolutional neural network (CNN)-based image super-resolutions are widely used as a high-quality image-enhancement technique. However, in general, they show little to no luminance isotropy. Thus, we propose two methods, “Luminance Inversion Training (LIT)” and “Luminance Inversion Averaging (LIA),” to improve the luminance isotropy of CNN-based image super-resolutions. Experimental results of 2× image magnification show that the average peak signal-to-noise ratio (PSNR) using Luminance Inversion Averaging is about 0.15-0.20dB higher than that for the conventional super-resolution.
Kazuya URAZOE
Kobe University
Nobutaka KUROKI
Kobe University
Yu KATO
Kobe University
Shinya OHTANI
Kobe University
Tetsuya HIROSE
Kobe University
Masahiro NUMA
Kobe University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Kazuya URAZOE, Nobutaka KUROKI, Yu KATO, Shinya OHTANI, Tetsuya HIROSE, Masahiro NUMA, "Improvement of Luminance Isotropy for Convolutional Neural Networks-Based Image Super-Resolution" in IEICE TRANSACTIONS on Fundamentals,
vol. E103-A, no. 7, pp. 955-958, July 2020, doi: 10.1587/transfun.2019EAL2168.
Abstract: Convolutional neural network (CNN)-based image super-resolutions are widely used as a high-quality image-enhancement technique. However, in general, they show little to no luminance isotropy. Thus, we propose two methods, “Luminance Inversion Training (LIT)” and “Luminance Inversion Averaging (LIA),” to improve the luminance isotropy of CNN-based image super-resolutions. Experimental results of 2× image magnification show that the average peak signal-to-noise ratio (PSNR) using Luminance Inversion Averaging is about 0.15-0.20dB higher than that for the conventional super-resolution.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.2019EAL2168/_p
Copy
@ARTICLE{e103-a_7_955,
author={Kazuya URAZOE, Nobutaka KUROKI, Yu KATO, Shinya OHTANI, Tetsuya HIROSE, Masahiro NUMA, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Improvement of Luminance Isotropy for Convolutional Neural Networks-Based Image Super-Resolution},
year={2020},
volume={E103-A},
number={7},
pages={955-958},
abstract={Convolutional neural network (CNN)-based image super-resolutions are widely used as a high-quality image-enhancement technique. However, in general, they show little to no luminance isotropy. Thus, we propose two methods, “Luminance Inversion Training (LIT)” and “Luminance Inversion Averaging (LIA),” to improve the luminance isotropy of CNN-based image super-resolutions. Experimental results of 2× image magnification show that the average peak signal-to-noise ratio (PSNR) using Luminance Inversion Averaging is about 0.15-0.20dB higher than that for the conventional super-resolution.},
keywords={},
doi={10.1587/transfun.2019EAL2168},
ISSN={1745-1337},
month={July},}
Copy
TY - JOUR
TI - Improvement of Luminance Isotropy for Convolutional Neural Networks-Based Image Super-Resolution
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 955
EP - 958
AU - Kazuya URAZOE
AU - Nobutaka KUROKI
AU - Yu KATO
AU - Shinya OHTANI
AU - Tetsuya HIROSE
AU - Masahiro NUMA
PY - 2020
DO - 10.1587/transfun.2019EAL2168
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E103-A
IS - 7
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - July 2020
AB - Convolutional neural network (CNN)-based image super-resolutions are widely used as a high-quality image-enhancement technique. However, in general, they show little to no luminance isotropy. Thus, we propose two methods, “Luminance Inversion Training (LIT)” and “Luminance Inversion Averaging (LIA),” to improve the luminance isotropy of CNN-based image super-resolutions. Experimental results of 2× image magnification show that the average peak signal-to-noise ratio (PSNR) using Luminance Inversion Averaging is about 0.15-0.20dB higher than that for the conventional super-resolution.
ER -