Understanding SSIM

Forskningsoutput: Bok/rapport › Rapport

Standard

Understanding SSIM. / Nilsson, Jim; Akenine-Möller, Tomas.

arXiv.org, 2020. 8 s.

Forskningsoutput: Bok/rapport › Rapport

Harvard

Nilsson, J & Akenine-Möller, T 2020, Understanding SSIM. arXiv.org. <https://arxiv.org/abs/2006.13846>

APA

Nilsson, J., & Akenine-Möller, T. (2020). Understanding SSIM. arXiv.org. https://arxiv.org/abs/2006.13846

CBE

Nilsson J, Akenine-Möller T 2020. Understanding SSIM. arXiv.org. 8 s.

MLA

Nilsson, Jim och Tomas Akenine-Möller Understanding SSIM arXiv.org. 2020.

Vancouver

Nilsson J, Akenine-Möller T. Understanding SSIM. arXiv.org, 2020. 8 s.

Author

Nilsson, Jim ; Akenine-Möller, Tomas. / Understanding SSIM. arXiv.org, 2020. 8 s.

RIS

TY - BOOK

T1 - Understanding SSIM

AU - Nilsson, Jim

AU - Akenine-Möller, Tomas

PY - 2020/6/29

Y1 - 2020/6/29

N2 - The use of the structural similarity index (SSIM) is widespread. For almost two decades, it has played a major role in image quality assessment in many different research disciplines. Clearly, its merits are indisputable in the research community. However, little deep scrutiny of this index has been performed. Contrary to popular belief, there are some interesting properties of SSIM that merit such scrutiny. In this paper, we analyze the mathematical factors of SSIM and show that it can generate results, in both synthetic and realistic use cases, that are unexpected, sometimes undefined, and nonintuitive. As a consequence, assessing image quality based on SSIM can lead to incorrect conclusions and using SSIM as a loss function for deep learning can guide neural network training in the wrong direction.

AB - The use of the structural similarity index (SSIM) is widespread. For almost two decades, it has played a major role in image quality assessment in many different research disciplines. Clearly, its merits are indisputable in the research community. However, little deep scrutiny of this index has been performed. Contrary to popular belief, there are some interesting properties of SSIM that merit such scrutiny. In this paper, we analyze the mathematical factors of SSIM and show that it can generate results, in both synthetic and realistic use cases, that are unexpected, sometimes undefined, and nonintuitive. As a consequence, assessing image quality based on SSIM can lead to incorrect conclusions and using SSIM as a loss function for deep learning can guide neural network training in the wrong direction.

UR - https://research.nvidia.com/publication/2020-07_Understanding-SSIM

M3 - Report

BT - Understanding SSIM

PB - arXiv.org

ER -