This paper proposes a new approach for filtering out impulse noise in digital color images. This work is of critical interest in biomedical imaging, visual tracking, etc. The conventional filtering methods operate by applying a noise reduction scheme, generally the vector median filtering approach and its variants, for the center pixel of a suitably chosen window that iteratively slides along the entire image. These methods consider the window in its entirety in the filtering process. This consideration, however, comprehends the noise within the noisy pixels in the filtering process and could prove detrimental to the overall output. The method proposed in this paper operates by clustering the pixels in the chosen window into two groups, one that corresponds to the pixel intensities that lie in the signal space and the other to those that lie in the noise space. The motivating rationale for this clustering scheme is to marginalize those pixels that lie in the noise space that seemingly do not contribute to the information in the image. The median filter is then applied to the pixels that contribute to the signal, in isolation of the color components, to filter out the impulse noise. Simulation results show that the proposed method outperforms conventional filtering methods in terms of noise reduction and structural similarity and thus validates the proposed approach. The proposed method is applied to analyze CT scans for improved diagnosis of SARS-COV-2 Covid19 disease. The method is also applied to a visual tracking example as a preprocessor.
The interest of this paper is in reduction of impulse noise in digital color images. The two main methods used for noise reduction in images are the mean and median filters. These techniques operate by replacing the test pixel in a chosen window by a new filtered pixel value. The window is made to iteratively slide across the entire image to reconstruct a new noise reduced image. The mean filters suffer from the effect of smoothing out color contrast and edges due to leveraging the unrepresentative pixels in the filtering process. The vector median filter and its variants overcome this problem by considering only the most representative pixel in the chosen window. The most representative pixel, i.e. the pixel that is of highest conformity to take the place of the test pixel, is determined by minimizing the aggregate distance from one pixel to every other pixel in the window. The problem in these median filtering approaches is that only one pixel is treated as representative of all the pixels in the chosen window. This conjecture could lead to information loss due to marginalizing other pixels that also are representative of the center pixel. In this paper, we propose a selective mean filtering process to overcome the said problem. The key idea here is to determine the most representative pixels in the window using the method of aggregate distances and then compute the mean of these pixels. This approach will perform better than the vector median filters as now a set of representative pixels are leveraged into the filtering process. Simulation results show that the proposed method performs better than the conventional vector median filtering methods in terms of noise reduction and structural similarity and thus validates the proposed approach. Moreover, the method is tested on real MRI scan images in successfully reducing impulse noise for improved medical diagnosis.
The focus of this paper is impulse noise reduction in digital color images. The most popular noise reduction schemes are the vector median filter and its many variants that operate by minimizing the aggregate distance from one pixel to every other pixel in a chosen window. This minimizing operation determines the most confirmative pixel based on its similarity to the chosen window and replaces the central pixel of the window with the determined one. The peer group filters, unlike the vector median filters, determine a set of pixels that are most confirmative to the window and then perform filtering over the determined set. Using a set of pixels in the filtering process rather than one pixel is more helpful as it takes into account the full information of all the pixels that seemingly contribute to the signal. Hence, the peer group filters are found to be more robust to noise. However, the peer group for each pixel is computed deterministically using thresholding schemes. A wrong choice of the threshold will easily impair the filtering performance. In this paper, we propose a peer group filtering approach using principles of Bayesian probability theory and clustering. Here, we present a method to compute the probability that a pixel value is clean (not corrupted by impulse noise) and then apply clustering on the probability measure to determine the peer group. The key benefit of this proposal is that the need for thresholding in peer group filtering is completely avoided. Simulation results show that the proposed method performs better than the conventional vector median and peer group filtering methods in terms of noise reduction and structural similarity, thus validating the proposed approach.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.