Image noise level estimation by searching for smooth patches with discrete cosine transform

Hayato Katase, Takuro Yamaguchi, Takanori Fujisawa, Masaaki Ikehara

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

When denoising an image, noise level is one of the most vital input parameters, because setting wrong noise level affects a result of denoising. Therefore, the noise level must be estimated accurately. In this paper, we propose a new accurate noise level estimation method based on the characteristic of discrete cosine transform (DCT) coefficients. This characteristic is that the high-frequency coefficients of the smooth patches can be assumed to zero. We select smooth patches from a distribution of the standard deviation of high-frequency coefficients in a noisy image, and a distribution function of the standard deviation of high-frequency coefficients in an only noise image. Moreover, we propose a method that the estimated noise level is calculated from high-frequency coefficients in a noisy image. The experiment results with many images demonstrate that the proposed method estimates the noise levels more accurately, in comparison with the conventional methods.

Original languageEnglish
Title of host publication2016 IEEE 59th International Midwest Symposium on Circuits and Systems, MWSCAS 2016
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781509009169
DOIs
Publication statusPublished - 2017 Mar 2
Event59th IEEE International Midwest Symposium on Circuits and Systems, MWSCAS 2016 - Abu Dhabi, United Arab Emirates
Duration: 2016 Oct 162016 Oct 19

Other

Other59th IEEE International Midwest Symposium on Circuits and Systems, MWSCAS 2016
CountryUnited Arab Emirates
CityAbu Dhabi
Period16/10/1616/10/19

Fingerprint

Image denoising
Discrete cosine transforms
Distribution functions
Experiments

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Electrical and Electronic Engineering

Cite this

Katase, H., Yamaguchi, T., Fujisawa, T., & Ikehara, M. (2017). Image noise level estimation by searching for smooth patches with discrete cosine transform. In 2016 IEEE 59th International Midwest Symposium on Circuits and Systems, MWSCAS 2016 [7870110] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/MWSCAS.2016.7870110

Image noise level estimation by searching for smooth patches with discrete cosine transform. / Katase, Hayato; Yamaguchi, Takuro; Fujisawa, Takanori; Ikehara, Masaaki.

2016 IEEE 59th International Midwest Symposium on Circuits and Systems, MWSCAS 2016. Institute of Electrical and Electronics Engineers Inc., 2017. 7870110.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Katase, H, Yamaguchi, T, Fujisawa, T & Ikehara, M 2017, Image noise level estimation by searching for smooth patches with discrete cosine transform. in 2016 IEEE 59th International Midwest Symposium on Circuits and Systems, MWSCAS 2016., 7870110, Institute of Electrical and Electronics Engineers Inc., 59th IEEE International Midwest Symposium on Circuits and Systems, MWSCAS 2016, Abu Dhabi, United Arab Emirates, 16/10/16. https://doi.org/10.1109/MWSCAS.2016.7870110
Katase H, Yamaguchi T, Fujisawa T, Ikehara M. Image noise level estimation by searching for smooth patches with discrete cosine transform. In 2016 IEEE 59th International Midwest Symposium on Circuits and Systems, MWSCAS 2016. Institute of Electrical and Electronics Engineers Inc. 2017. 7870110 https://doi.org/10.1109/MWSCAS.2016.7870110
Katase, Hayato ; Yamaguchi, Takuro ; Fujisawa, Takanori ; Ikehara, Masaaki. / Image noise level estimation by searching for smooth patches with discrete cosine transform. 2016 IEEE 59th International Midwest Symposium on Circuits and Systems, MWSCAS 2016. Institute of Electrical and Electronics Engineers Inc., 2017.
@inproceedings{2eb32e997f7c4114a570aff9f4fb7c58,
title = "Image noise level estimation by searching for smooth patches with discrete cosine transform",
abstract = "When denoising an image, noise level is one of the most vital input parameters, because setting wrong noise level affects a result of denoising. Therefore, the noise level must be estimated accurately. In this paper, we propose a new accurate noise level estimation method based on the characteristic of discrete cosine transform (DCT) coefficients. This characteristic is that the high-frequency coefficients of the smooth patches can be assumed to zero. We select smooth patches from a distribution of the standard deviation of high-frequency coefficients in a noisy image, and a distribution function of the standard deviation of high-frequency coefficients in an only noise image. Moreover, we propose a method that the estimated noise level is calculated from high-frequency coefficients in a noisy image. The experiment results with many images demonstrate that the proposed method estimates the noise levels more accurately, in comparison with the conventional methods.",
author = "Hayato Katase and Takuro Yamaguchi and Takanori Fujisawa and Masaaki Ikehara",
year = "2017",
month = "3",
day = "2",
doi = "10.1109/MWSCAS.2016.7870110",
language = "English",
booktitle = "2016 IEEE 59th International Midwest Symposium on Circuits and Systems, MWSCAS 2016",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Image noise level estimation by searching for smooth patches with discrete cosine transform

AU - Katase, Hayato

AU - Yamaguchi, Takuro

AU - Fujisawa, Takanori

AU - Ikehara, Masaaki

PY - 2017/3/2

Y1 - 2017/3/2

N2 - When denoising an image, noise level is one of the most vital input parameters, because setting wrong noise level affects a result of denoising. Therefore, the noise level must be estimated accurately. In this paper, we propose a new accurate noise level estimation method based on the characteristic of discrete cosine transform (DCT) coefficients. This characteristic is that the high-frequency coefficients of the smooth patches can be assumed to zero. We select smooth patches from a distribution of the standard deviation of high-frequency coefficients in a noisy image, and a distribution function of the standard deviation of high-frequency coefficients in an only noise image. Moreover, we propose a method that the estimated noise level is calculated from high-frequency coefficients in a noisy image. The experiment results with many images demonstrate that the proposed method estimates the noise levels more accurately, in comparison with the conventional methods.

AB - When denoising an image, noise level is one of the most vital input parameters, because setting wrong noise level affects a result of denoising. Therefore, the noise level must be estimated accurately. In this paper, we propose a new accurate noise level estimation method based on the characteristic of discrete cosine transform (DCT) coefficients. This characteristic is that the high-frequency coefficients of the smooth patches can be assumed to zero. We select smooth patches from a distribution of the standard deviation of high-frequency coefficients in a noisy image, and a distribution function of the standard deviation of high-frequency coefficients in an only noise image. Moreover, we propose a method that the estimated noise level is calculated from high-frequency coefficients in a noisy image. The experiment results with many images demonstrate that the proposed method estimates the noise levels more accurately, in comparison with the conventional methods.

UR - http://www.scopus.com/inward/record.url?scp=85015907049&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85015907049&partnerID=8YFLogxK

U2 - 10.1109/MWSCAS.2016.7870110

DO - 10.1109/MWSCAS.2016.7870110

M3 - Conference contribution

BT - 2016 IEEE 59th International Midwest Symposium on Circuits and Systems, MWSCAS 2016

PB - Institute of Electrical and Electronics Engineers Inc.

ER -