Multispectral image denoising with optimized vector non-local mean filter

Ahmed Ben Said, Rachid Hadjidj, Kamal Eddine Melkemi, Sebti Foufou

Research output: Contribution to journalArticle

Abstract

Nowadays, many applications rely on images of high quality to ensure good performance in conducting their tasks. However, noise goes against this objective as it is an unavoidable issue in most applications. Therefore, it is essential to develop techniques to attenuate the impact of noise, while maintaining the integrity of relevant information in images. We propose in this work to extend the application of the Non-Local Means filter (NLM) to the vector case and apply it for denoising multispectral images. The objective is to benefit from the additional information brought by multispectral imaging systems. The NLM filter exploits the redundancy of information in an image to remove noise. A restored pixel is a weighted average of all pixels in the image. In our contribution, we propose an optimization framework where we dynamically fine tune the NLM filter parameters and attenuate its computational complexity by considering only pixels which are most similar to each other in computing a restored pixel. Filter parameters are optimized using Stein's Unbiased Risk Estimator (SURE) rather than using ad hoc means. Experiments have been conducted on multispectral images corrupted with additive white Gaussian noise. PSNR and similarity comparison with other approaches are provided to illustrate the efficiency of our approach in terms of both denoising performance and computation complexity.

Original languageEnglish (US)
Pages (from-to)115-126
Number of pages12
JournalDigital Signal Processing: A Review Journal
Volume58
DOIs
StatePublished - Nov 1 2016

Fingerprint

Image denoising
Pixels
Imaging systems
Redundancy
Computational complexity
Experiments

Keywords

  • Image denoising
  • Multispectral image
  • Stein's unbiased risk estimator
  • Vector non-local mean filter

ASJC Scopus subject areas

  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Multispectral image denoising with optimized vector non-local mean filter. / Ben Said, Ahmed; Hadjidj, Rachid; Eddine Melkemi, Kamal; Foufou, Sebti.

In: Digital Signal Processing: A Review Journal, Vol. 58, 01.11.2016, p. 115-126.

Research output: Contribution to journalArticle

Ben Said, Ahmed ; Hadjidj, Rachid ; Eddine Melkemi, Kamal ; Foufou, Sebti. / Multispectral image denoising with optimized vector non-local mean filter. In: Digital Signal Processing: A Review Journal. 2016 ; Vol. 58. pp. 115-126.
@article{198c4379243a4dc094bb893dcf7e56d4,
title = "Multispectral image denoising with optimized vector non-local mean filter",
abstract = "Nowadays, many applications rely on images of high quality to ensure good performance in conducting their tasks. However, noise goes against this objective as it is an unavoidable issue in most applications. Therefore, it is essential to develop techniques to attenuate the impact of noise, while maintaining the integrity of relevant information in images. We propose in this work to extend the application of the Non-Local Means filter (NLM) to the vector case and apply it for denoising multispectral images. The objective is to benefit from the additional information brought by multispectral imaging systems. The NLM filter exploits the redundancy of information in an image to remove noise. A restored pixel is a weighted average of all pixels in the image. In our contribution, we propose an optimization framework where we dynamically fine tune the NLM filter parameters and attenuate its computational complexity by considering only pixels which are most similar to each other in computing a restored pixel. Filter parameters are optimized using Stein's Unbiased Risk Estimator (SURE) rather than using ad hoc means. Experiments have been conducted on multispectral images corrupted with additive white Gaussian noise. PSNR and similarity comparison with other approaches are provided to illustrate the efficiency of our approach in terms of both denoising performance and computation complexity.",
keywords = "Image denoising, Multispectral image, Stein's unbiased risk estimator, Vector non-local mean filter",
author = "{Ben Said}, Ahmed and Rachid Hadjidj and {Eddine Melkemi}, Kamal and Sebti Foufou",
year = "2016",
month = "11",
day = "1",
doi = "10.1016/j.dsp.2016.07.017",
language = "English (US)",
volume = "58",
pages = "115--126",
journal = "Digital Signal Processing: A Review Journal",
issn = "1051-2004",
publisher = "Elsevier Inc.",

}

TY - JOUR

T1 - Multispectral image denoising with optimized vector non-local mean filter

AU - Ben Said, Ahmed

AU - Hadjidj, Rachid

AU - Eddine Melkemi, Kamal

AU - Foufou, Sebti

PY - 2016/11/1

Y1 - 2016/11/1

N2 - Nowadays, many applications rely on images of high quality to ensure good performance in conducting their tasks. However, noise goes against this objective as it is an unavoidable issue in most applications. Therefore, it is essential to develop techniques to attenuate the impact of noise, while maintaining the integrity of relevant information in images. We propose in this work to extend the application of the Non-Local Means filter (NLM) to the vector case and apply it for denoising multispectral images. The objective is to benefit from the additional information brought by multispectral imaging systems. The NLM filter exploits the redundancy of information in an image to remove noise. A restored pixel is a weighted average of all pixels in the image. In our contribution, we propose an optimization framework where we dynamically fine tune the NLM filter parameters and attenuate its computational complexity by considering only pixels which are most similar to each other in computing a restored pixel. Filter parameters are optimized using Stein's Unbiased Risk Estimator (SURE) rather than using ad hoc means. Experiments have been conducted on multispectral images corrupted with additive white Gaussian noise. PSNR and similarity comparison with other approaches are provided to illustrate the efficiency of our approach in terms of both denoising performance and computation complexity.

AB - Nowadays, many applications rely on images of high quality to ensure good performance in conducting their tasks. However, noise goes against this objective as it is an unavoidable issue in most applications. Therefore, it is essential to develop techniques to attenuate the impact of noise, while maintaining the integrity of relevant information in images. We propose in this work to extend the application of the Non-Local Means filter (NLM) to the vector case and apply it for denoising multispectral images. The objective is to benefit from the additional information brought by multispectral imaging systems. The NLM filter exploits the redundancy of information in an image to remove noise. A restored pixel is a weighted average of all pixels in the image. In our contribution, we propose an optimization framework where we dynamically fine tune the NLM filter parameters and attenuate its computational complexity by considering only pixels which are most similar to each other in computing a restored pixel. Filter parameters are optimized using Stein's Unbiased Risk Estimator (SURE) rather than using ad hoc means. Experiments have been conducted on multispectral images corrupted with additive white Gaussian noise. PSNR and similarity comparison with other approaches are provided to illustrate the efficiency of our approach in terms of both denoising performance and computation complexity.

KW - Image denoising

KW - Multispectral image

KW - Stein's unbiased risk estimator

KW - Vector non-local mean filter

UR - http://www.scopus.com/inward/record.url?scp=84984673952&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84984673952&partnerID=8YFLogxK

U2 - 10.1016/j.dsp.2016.07.017

DO - 10.1016/j.dsp.2016.07.017

M3 - Article

AN - SCOPUS:84984673952

VL - 58

SP - 115

EP - 126

JO - Digital Signal Processing: A Review Journal

JF - Digital Signal Processing: A Review Journal

SN - 1051-2004

ER -