Novel method for removal of multichannel impulse noise based on half-space deepest location

Djordje Baljozovic; Branko Kovacevic; Aleksandra Baljozovic

doi:10.1117/1.JEI.21.1.013025

16 April 2012 Novel method for removal of multichannel impulse noise based on half-space deepest location

Djordje Baljozovic, Branko Kovacevic, Aleksandra Baljozovic

Author Affiliations +

Journal of Electronic Imaging, Vol. 21, Issue 1, 013025 (April 2012). https://doi.org/10.1117/1.JEI.21.1.013025

Abstract

A novel method is proposed for removing impulse (random-valued and salt-and-pepper) noise from multichannel digital images based on an adapted version of the DEEPLOC algorithm introduced by A. Struyf and P. J. Rousseeuw for calculation of approximate half-space (Tukey's) deepest location (median) in multivariate case. Due to its intrinsic multivariate/multidimensional nature, the proposed method eliminates the noise simultaneously on all channels without their separation, which preserves the spectral correlation between channels in a multichannel image. Denoising results of this new nonlinear spatial domain filter applied to benchmark images outperform currently used state-of-the-art filters for impulse noise removal from multichannel images in terms of both objective effectiveness criteria [peak-signal-to-noise-ratio (PSNR), mean absolute error (MAE), and normalized color distance (NCD)] and visual quality. The proposed filter successfully preserves the edges and fine image details, and is very effective for removal of medium and heavy multichannel impulse noise.

Citation Download Citation

Djordje Baljozovic, Branko Kovacevic, and Aleksandra Baljozovic "Novel method for removal of multichannel impulse noise based on half-space deepest location," Journal of Electronic Imaging 21(1), 013025 (16 April 2012). https://doi.org/10.1117/1.JEI.21.1.013025

Published: 16 April 2012

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available