The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Este artigo aborda super-resolução de imagem única (SR) baseada em redes neurais convolucionais (CNNs). Sabe-se que a recuperação de componentes de alta frequência em imagens SR de saída de CNNs aprendidas pelos mínimos erros quadrados ou mínimos erros absolutos é insuficiente. Para gerar componentes realistas de alta frequência, métodos SR usando redes adversárias generativas (GANs), compostas por um gerador e um discriminador, São desenvolvidos. No entanto, quando o gerador tenta induzir o erro de julgamento do discriminador, não apenas componentes realistas de alta frequência, mas também alguns artefatos são gerados e índices objetivos como PSNR diminuem. Para reduzir os artefatos nos métodos SR baseados em GAN, consideramos o conjunto de todas as imagens SR cujos erros quadrados entre os resultados de redução de escala e a imagem de entrada estão dentro de um determinado intervalo, e propomos aplicar o projeção métrica para isso conjunto consistente nas camadas de saída dos geradores. A técnica proposta garante a consistência entre as imagens SR de saída e as imagens de entrada, e os geradores com a projeção proposta podem gerar componentes de alta frequência com poucos artefatos, mantendo os de baixa frequência adequados ao nível de ruído conhecido. Experimentos numéricos mostram que a técnica proposta reduz artefatos incluídos nas imagens SR originais de um método SR baseado em GAN, ao mesmo tempo que gera componentes realistas de alta frequência com melhores valores de PSNR em ambos sem ruído e ruidoso situações. Como a técnica proposta pode ser integrada em vários geradores se o processo de downscaling for conhecido, podemos dar consistência aos métodos existentes com as imagens de entrada sem degradar o desempenho de outros SR.
Hiroya YAMAMOTO
Ritsumeikan University
Daichi KITAHARA
Ritsumeikan University
Hiroki KURODA
Ritsumeikan University
Akira HIRABAYASHI
Ritsumeikan 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.
Copiar
Hiroya YAMAMOTO, Daichi KITAHARA, Hiroki KURODA, Akira HIRABAYASHI, "Image Super-Resolution via Generative Adversarial Networks Using Metric Projections onto Consistent Sets for Low-Resolution Inputs" in IEICE TRANSACTIONS on Fundamentals,
vol. E105-A, no. 4, pp. 704-718, April 2022, doi: 10.1587/transfun.2021EAP1038.
Abstract: This paper addresses single image super-resolution (SR) based on convolutional neural networks (CNNs). It is known that recovery of high-frequency components in output SR images of CNNs learned by the least square errors or least absolute errors is insufficient. To generate realistic high-frequency components, SR methods using generative adversarial networks (GANs), composed of one generator and one discriminator, are developed. However, when the generator tries to induce the discriminator's misjudgment, not only realistic high-frequency components but also some artifacts are generated, and objective indices such as PSNR decrease. To reduce the artifacts in the GAN-based SR methods, we consider the set of all SR images whose square errors between downscaling results and the input image are within a certain range, and propose to apply the metric projection onto this consistent set in the output layers of the generators. The proposed technique guarantees the consistency between output SR images and input images, and the generators with the proposed projection can generate high-frequency components with few artifacts while keeping low-frequency ones as appropriate for the known noise level. Numerical experiments show that the proposed technique reduces artifacts included in the original SR images of a GAN-based SR method while generating realistic high-frequency components with better PSNR values in both noise-free and noisy situations. Since the proposed technique can be integrated into various generators if the downscaling process is known, we can give the consistency to existing methods with the input images without degrading other SR performance.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.2021EAP1038/_p
Copiar
@ARTICLE{e105-a_4_704,
author={Hiroya YAMAMOTO, Daichi KITAHARA, Hiroki KURODA, Akira HIRABAYASHI, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Image Super-Resolution via Generative Adversarial Networks Using Metric Projections onto Consistent Sets for Low-Resolution Inputs},
year={2022},
volume={E105-A},
number={4},
pages={704-718},
abstract={This paper addresses single image super-resolution (SR) based on convolutional neural networks (CNNs). It is known that recovery of high-frequency components in output SR images of CNNs learned by the least square errors or least absolute errors is insufficient. To generate realistic high-frequency components, SR methods using generative adversarial networks (GANs), composed of one generator and one discriminator, are developed. However, when the generator tries to induce the discriminator's misjudgment, not only realistic high-frequency components but also some artifacts are generated, and objective indices such as PSNR decrease. To reduce the artifacts in the GAN-based SR methods, we consider the set of all SR images whose square errors between downscaling results and the input image are within a certain range, and propose to apply the metric projection onto this consistent set in the output layers of the generators. The proposed technique guarantees the consistency between output SR images and input images, and the generators with the proposed projection can generate high-frequency components with few artifacts while keeping low-frequency ones as appropriate for the known noise level. Numerical experiments show that the proposed technique reduces artifacts included in the original SR images of a GAN-based SR method while generating realistic high-frequency components with better PSNR values in both noise-free and noisy situations. Since the proposed technique can be integrated into various generators if the downscaling process is known, we can give the consistency to existing methods with the input images without degrading other SR performance.},
keywords={},
doi={10.1587/transfun.2021EAP1038},
ISSN={1745-1337},
month={April},}
Copiar
TY - JOUR
TI - Image Super-Resolution via Generative Adversarial Networks Using Metric Projections onto Consistent Sets for Low-Resolution Inputs
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 704
EP - 718
AU - Hiroya YAMAMOTO
AU - Daichi KITAHARA
AU - Hiroki KURODA
AU - Akira HIRABAYASHI
PY - 2022
DO - 10.1587/transfun.2021EAP1038
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E105-A
IS - 4
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - April 2022
AB - This paper addresses single image super-resolution (SR) based on convolutional neural networks (CNNs). It is known that recovery of high-frequency components in output SR images of CNNs learned by the least square errors or least absolute errors is insufficient. To generate realistic high-frequency components, SR methods using generative adversarial networks (GANs), composed of one generator and one discriminator, are developed. However, when the generator tries to induce the discriminator's misjudgment, not only realistic high-frequency components but also some artifacts are generated, and objective indices such as PSNR decrease. To reduce the artifacts in the GAN-based SR methods, we consider the set of all SR images whose square errors between downscaling results and the input image are within a certain range, and propose to apply the metric projection onto this consistent set in the output layers of the generators. The proposed technique guarantees the consistency between output SR images and input images, and the generators with the proposed projection can generate high-frequency components with few artifacts while keeping low-frequency ones as appropriate for the known noise level. Numerical experiments show that the proposed technique reduces artifacts included in the original SR images of a GAN-based SR method while generating realistic high-frequency components with better PSNR values in both noise-free and noisy situations. Since the proposed technique can be integrated into various generators if the downscaling process is known, we can give the consistency to existing methods with the input images without degrading other SR performance.
ER -