Underwater image denoising via AMP with renyi entropy

Jihua Lu; Lei Yan; Huan Wang; Lihui Feng; Xinbin Li

doi:10.1145/3366486.3366531

Underwater image denoising via AMP with renyi entropy

Jihua Lu
, Lei Yan
, Huan Wang
, Lihui Feng
, Xinbin Li

集成电路与电子学院

Yanshan University
Beijing Institute of Technology

科研成果: 书/报告/会议事项章节 › 会议稿件 › 同行评审

摘要

Approximate message passing (AMP) in wavelet domain has been widely used in image denoising. However, the AMP-based denoising method is not robust to the 'shot noise', which is dynamic and short-period in the wavelet domain. First, a preprocessing filter is designed with Entropy to combat the restlessness of noise component in spatial domain. Then, a shot noise filter is designed via AMP with Renyi Entropy in wavelet domain. By using the two-step denoising method, we can 'individualized' denoise both in spatial and wavelet domains. The experimental results show that, with preprocessing filtering, our denoising method outperform the design via AMP with wavelet thresholds.

源语言	英语
主期刊名	Proceedings of the International Conference on Underwater Networks and Systems, WUWNET 2019
出版商	Association for Computing Machinery
ISBN（电子版）	9781450377409
DOI	https://doi.org/10.1145/3366486.3366531
出版状态	已出版 - 23 10月 2019
活动	2019 International Conference on Underwater Networks and Systems, WUWNET 2019 - Atlanta, 美国期限: 23 10月 2019 → 25 10月 2019

丛书

姓名	ACM International Conference Proceeding Series

会议

会议	2019 International Conference on Underwater Networks and Systems, WUWNET 2019
国家/地区	美国
市	Atlanta
时期	23/10/19 → 25/10/19

访问文件

10.1145/3366486.3366531

其它文件与链接

链接到 Scopus 的出版物

引用此

@inproceedings{e62cfd1e84bc4eb18024bbfd38ab7814,

title = "Underwater image denoising via AMP with renyi entropy",

abstract = "Approximate message passing (AMP) in wavelet domain has been widely used in image denoising. However, the AMP-based denoising method is not robust to the 'shot noise', which is dynamic and short-period in the wavelet domain. First, a preprocessing filter is designed with Entropy to combat the restlessness of noise component in spatial domain. Then, a shot noise filter is designed via AMP with Renyi Entropy in wavelet domain. By using the two-step denoising method, we can 'individualized' denoise both in spatial and wavelet domains. The experimental results show that, with preprocessing filtering, our denoising method outperform the design via AMP with wavelet thresholds.",

keywords = "Approximate message passing, Renyi Entropy, Shot noise, Underwater image denoising, Wavelet transform",

author = "Jihua Lu and Lei Yan and Huan Wang and Lihui Feng and Xinbin Li",

note = "Publisher Copyright: {\textcopyright} 2019 Authors.; 2019 International Conference on Underwater Networks and Systems, WUWNET 2019 ; Conference date: 23-10-2019 Through 25-10-2019",

year = "2019",

month = oct,

day = "23",

doi = "10.1145/3366486.3366531",

language = "English",

series = "ACM International Conference Proceeding Series",

publisher = "Association for Computing Machinery",

booktitle = "Proceedings of the International Conference on Underwater Networks and Systems, WUWNET 2019",

}

Lu, J, Yan, L, Wang, H, Feng, L & Li, X 2019, Underwater image denoising via AMP with renyi entropy. 在 Proceedings of the International Conference on Underwater Networks and Systems, WUWNET 2019., a6, ACM International Conference Proceeding Series, Association for Computing Machinery, 2019 International Conference on Underwater Networks and Systems, WUWNET 2019, Atlanta, 美国, 23/10/19. https://doi.org/10.1145/3366486.3366531

TY - GEN

T1 - Underwater image denoising via AMP with renyi entropy

AU - Lu, Jihua

AU - Yan, Lei

AU - Wang, Huan

AU - Feng, Lihui

AU - Li, Xinbin

PY - 2019/10/23

Y1 - 2019/10/23

N2 - Approximate message passing (AMP) in wavelet domain has been widely used in image denoising. However, the AMP-based denoising method is not robust to the 'shot noise', which is dynamic and short-period in the wavelet domain. First, a preprocessing filter is designed with Entropy to combat the restlessness of noise component in spatial domain. Then, a shot noise filter is designed via AMP with Renyi Entropy in wavelet domain. By using the two-step denoising method, we can 'individualized' denoise both in spatial and wavelet domains. The experimental results show that, with preprocessing filtering, our denoising method outperform the design via AMP with wavelet thresholds.

AB - Approximate message passing (AMP) in wavelet domain has been widely used in image denoising. However, the AMP-based denoising method is not robust to the 'shot noise', which is dynamic and short-period in the wavelet domain. First, a preprocessing filter is designed with Entropy to combat the restlessness of noise component in spatial domain. Then, a shot noise filter is designed via AMP with Renyi Entropy in wavelet domain. By using the two-step denoising method, we can 'individualized' denoise both in spatial and wavelet domains. The experimental results show that, with preprocessing filtering, our denoising method outperform the design via AMP with wavelet thresholds.

KW - Approximate message passing

KW - Renyi Entropy

KW - Shot noise

KW - Underwater image denoising

KW - Wavelet transform

UR - https://www.scopus.com/pages/publications/85081050399

U2 - 10.1145/3366486.3366531

DO - 10.1145/3366486.3366531

M3 - Conference contribution

AN - SCOPUS:85081050399

T3 - ACM International Conference Proceeding Series

BT - Proceedings of the International Conference on Underwater Networks and Systems, WUWNET 2019

PB - Association for Computing Machinery

T2 - 2019 International Conference on Underwater Networks and Systems, WUWNET 2019

Y2 - 23 October 2019 through 25 October 2019

ER -

Underwater image denoising via AMP with renyi entropy

摘要

丛书

会议

访问文件

其它文件与链接

学术指纹

引用此