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
A avaliação da qualidade da imagem (IQA) é uma métrica fundamental para tarefas de processamento de imagens (por exemplo, compressão). Com IQAs de referência completa, foram utilizados IQAs tradicionais, como PSNR e SSIM. Recentemente, também foram utilizados IQAs baseados em redes neurais profundas (Deep IQAs), como LPIPS e DISTS. Sabe-se que o dimensionamento da imagem é inconsistente entre IQAs profundos, pois alguns realizam redução de escala como pré-processamento, enquanto outros usam o tamanho original da imagem. Neste artigo, mostramos que a escala da imagem é um fator influente que afeta o desempenho do IQA profundo. Avaliamos de forma abrangente quatro IQAs profundos nos mesmos cinco conjuntos de dados, e os resultados experimentais mostram que a escala da imagem influencia significativamente o desempenho do IQA. Descobrimos que a escala de imagem mais apropriada muitas vezes não é o tamanho padrão nem o tamanho original, e a escolha difere dependendo dos métodos e conjuntos de dados usados. Visualizamos a estabilidade e descobrimos que o PieAPP é o mais estável entre os quatro IQAs profundos.
Koki TSUBOTA
The University of Tokyo
Hiroaki AKUTSU
Hitachi, Ltd.
Kiyoharu AIZAWA
The University of Tokyo
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
Koki TSUBOTA, Hiroaki AKUTSU, Kiyoharu AIZAWA, "Evaluating the Stability of Deep Image Quality Assessment with Respect to Image Scaling" in IEICE TRANSACTIONS on Information,
vol. E105-D, no. 10, pp. 1829-1833, October 2022, doi: 10.1587/transinf.2022EDL8025.
Abstract: Image quality assessment (IQA) is a fundamental metric for image processing tasks (e.g., compression). With full-reference IQAs, traditional IQAs, such as PSNR and SSIM, have been used. Recently, IQAs based on deep neural networks (deep IQAs), such as LPIPS and DISTS, have also been used. It is known that image scaling is inconsistent among deep IQAs, as some perform down-scaling as pre-processing, whereas others instead use the original image size. In this paper, we show that the image scale is an influential factor that affects deep IQA performance. We comprehensively evaluate four deep IQAs on the same five datasets, and the experimental results show that image scale significantly influences IQA performance. We found that the most appropriate image scale is often neither the default nor the original size, and the choice differs depending on the methods and datasets used. We visualized the stability and found that PieAPP is the most stable among the four deep IQAs.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2022EDL8025/_p
Copiar
@ARTICLE{e105-d_10_1829,
author={Koki TSUBOTA, Hiroaki AKUTSU, Kiyoharu AIZAWA, },
journal={IEICE TRANSACTIONS on Information},
title={Evaluating the Stability of Deep Image Quality Assessment with Respect to Image Scaling},
year={2022},
volume={E105-D},
number={10},
pages={1829-1833},
abstract={Image quality assessment (IQA) is a fundamental metric for image processing tasks (e.g., compression). With full-reference IQAs, traditional IQAs, such as PSNR and SSIM, have been used. Recently, IQAs based on deep neural networks (deep IQAs), such as LPIPS and DISTS, have also been used. It is known that image scaling is inconsistent among deep IQAs, as some perform down-scaling as pre-processing, whereas others instead use the original image size. In this paper, we show that the image scale is an influential factor that affects deep IQA performance. We comprehensively evaluate four deep IQAs on the same five datasets, and the experimental results show that image scale significantly influences IQA performance. We found that the most appropriate image scale is often neither the default nor the original size, and the choice differs depending on the methods and datasets used. We visualized the stability and found that PieAPP is the most stable among the four deep IQAs.},
keywords={},
doi={10.1587/transinf.2022EDL8025},
ISSN={1745-1361},
month={October},}
Copiar
TY - JOUR
TI - Evaluating the Stability of Deep Image Quality Assessment with Respect to Image Scaling
T2 - IEICE TRANSACTIONS on Information
SP - 1829
EP - 1833
AU - Koki TSUBOTA
AU - Hiroaki AKUTSU
AU - Kiyoharu AIZAWA
PY - 2022
DO - 10.1587/transinf.2022EDL8025
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E105-D
IS - 10
JA - IEICE TRANSACTIONS on Information
Y1 - October 2022
AB - Image quality assessment (IQA) is a fundamental metric for image processing tasks (e.g., compression). With full-reference IQAs, traditional IQAs, such as PSNR and SSIM, have been used. Recently, IQAs based on deep neural networks (deep IQAs), such as LPIPS and DISTS, have also been used. It is known that image scaling is inconsistent among deep IQAs, as some perform down-scaling as pre-processing, whereas others instead use the original image size. In this paper, we show that the image scale is an influential factor that affects deep IQA performance. We comprehensively evaluate four deep IQAs on the same five datasets, and the experimental results show that image scale significantly influences IQA performance. We found that the most appropriate image scale is often neither the default nor the original size, and the choice differs depending on the methods and datasets used. We visualized the stability and found that PieAPP is the most stable among the four deep IQAs.
ER -