Restoration of highly salt-and-pepper-noise-corrupted images using novel adaptive trimmed median filter
- PDF / 1,299,690 Bytes
- 9 Pages / 595.276 x 790.866 pts Page_size
- 33 Downloads / 170 Views
ORIGINAL PAPER
Restoration of highly salt-and-pepper-noise-corrupted images using novel adaptive trimmed median filter Bharat Garg1 Received: 20 July 2019 / Revised: 1 March 2020 / Accepted: 17 April 2020 © Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract The paper presents a novel adaptive trimmed median (ATM) filter to remove salt-and-pepper (SAP) noise of high noise density (ND). The proposed filter computes median of trimmed window of adaptive size containing noise-free pixels (NFP) for ND up medium range while performs new interpolation-based procedure at high ND. Further, for the rare scenarios especially at the boundary where denoising using interpolation is not good enough, the proposed filter denoises the candidate pixel with the help of nearest processed pixels. The proposed ATM filter effectively suppresses SAP noise because denoising mostly utilizes original non-noisy pixels. The proposed algorithm is evaluated for varying ND (10–90%) with different benchmark images (greyscale and coloured) over the existing approaches. The proposed ATM filter on an average provides 1.59 dB and 0.37 dB higher PSNR values on the greyscale and color images, respectively. Keywords Median filter · Interpolation · Salt-and-pepper noise · Image processing
1 Introduction The images are inevitably contaminated by the impulse noise during image acquisition, transmission and/or storage due to malfunctioning of camera sensors, noise in transmission channel and/or fault in memory, respectively. In the impulse noise, salt-and-pepper (SAP) noise significantly degrades image quality; therefore, nonlinear filters are presented to retrieve the noise-free image. Among the several nonlinear filters, the median filter (MF) is more popular for the removal of SAP noise without destroying image information. However, the standard MF (SMF) is only able to recover image corrupted with low noise density (N D) [1]. At higher ND (N D > 50%), the SMF does not get sufficient noise-free pixels (NFPs) within local window to restore noisy-pixel and therefore fails to recover original image. The adaptive MF (AMF) [2] increases the window size with ND to remove corrupted pixel. However, it exhibits blurring effects at higher noise density. The modified switching MFs [3,4] take decision on the noisy pixel on the basis of the pre-defined threshold. The prime limitation of these filters is the implementation of method that provides correct deci-
B 1
Bharat Garg [email protected] Thapar Institute of Engineering and Technology, Patiala, India
sion. An improved tolerance-based selective arithmetic mean (ITSAM) filter [5] detects the ND, and if it is found above a given threshold, it restores noisy-pixel by the mean of current window (CW), else leaves the pixel unaltered by considering it as NFP. A decision-based algorithm for median (DBAM) filter [6] performs the sorting in CW horizontally, vertically and diagonally and takes the centre pixel as denoised pixel. If the pixel is still noisy, the algorithm restores the pixel by previou
Data Loading...
