A novel decision-based adaptive feedback median filter for high density impulse noise suppression
- PDF / 1,967,055 Bytes
- 23 Pages / 439.37 x 666.142 pts Page_size
- 50 Downloads / 189 Views
A novel decision-based adaptive feedback median filter for high density impulse noise suppression Kamarujjaman 1
& Mausumi Maitra
1
& Susanta Chakraborty
2
Received: 13 July 2019 / Revised: 20 July 2020 / Accepted: 28 July 2020 # Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract
The qualitative performances of the digital image processing methods are degraded due to the presence of impulse noise. The conventional median filter and its advanced versions somehow manage to remove the noise from image but cannot preserve the image details. In this paper, a novel decision based adaptive feedback median filter is proposed to suppress the high density noise and preserve the details of the image. The proposed method detects the corrupted or noisy pixels by analyzing the neighbours in a decisive manner, which is a challenging task for the different types of images and noise. It predicts a local threshold by analyzing the neighbours to decide the adaptive nature of the feedback median filter. The feedback mechanism is adapted to enhance the qualitative results. Various types of images and noise densities have been used to evaluate the performance of the proposed method. The qualitative and quantitative performances have been measured in terms of Peak Signal-to-Noise Ratio, Image Enhancement Factor and Structural Similarity Index. The experimental results show that the qualitative and quantitative performances are superior over existing methods and the computational time is comparable as well. Keywords Decision based median filter . Image denoising . Impulse noise . Peak signal-to-noise ratio . Salt and pepper
1 Introduction Digital images are distorted due to impulse noise, which is the result of malfunction of the camera sensor and/or missing memory in the imaging hardware [6]. The corrupted images can * Mausumi Maitra [email protected] Kamarujjaman [email protected] Susanta Chakraborty [email protected] Extended author information available on the last page of the article
Multimedia Tools and Applications
be retrieved simply using non-linear median filters [17], adaptive median filter [18], and adaptive center-weighted median filter (ACWMF) [7] but by observing the results it has been seen that with increasing density of impulse noise the results of the mentioned methods are degraded. The switching based methods are developed in [6, 11–13, 19, 21, 22, 28, 29, 35] to suppress the impulse noise even for higher density levels in images. In this type of filters two steps are incorporated, one is for noise detection and another one for noise suppression [27]. The mean filtering, block-matching filtering and the updated versions of these filters [3–5, 8, 9] have been proposed to restore the corrupted images using global information. The generative classification methods based on non local mean (NLM) have been reported in a recent literature [16] to suppress the high density impulse noise. In [25] difference median filter has been proposed to remove the impulse noise by detecting th
Data Loading...
