Theses / Dissertations >
College of Engineering >
Please use this identifier to cite or link to this item:
|Title: ||Improved median filtering algorithm for the reduction of impulse noise in corrupted 2d greyscale images|
|Authors: ||Weyori, Benjamin Asubam|
|Issue Date: ||29-Nov-2011|
|Abstract: ||Digital images are often corrupted by Impulse noise due to errors generated in noisy sensor, errors that occur in the process of converting signals from analog-to-digital and also errors that are generated in the communication channels. This error that occurs inevitably alters some of the pixels intensity while some of the pixels remain unchanged. In order to remove impulse noise and enhance the affected image quality, the median filter has been studied and a method based on an improved median filtering algorithm has been proposed. This method removes or effectively suppresses the impulse noise in the image whiles preserving the image edges information and enhancing the image quality.
The proposed method is a spatial domain approach and uses the overlapping window to filter the signal based on the selection of an effective median per window. The approach chosen in this work is based on a functional level 2n +1 window that makes the selection of the normal median easier, since the number of elements in the window is odd. The median so chosen is confirmed as the effective median or, where the median is an impulse a more representative value is sought and used as the effective median. The improved median filtering algorithms uses the median switching technique to compute an effective median when the active median of the window is an impulse.
The performance of the proposed effective median filter has been evaluated in MATLAB using a 3 × 3 fixed window for simulations on an image that has been subjected to various degrees of corruption with impulse noise. The results demonstrate the effectiveness of the proposed algorithm vis-à-vis the standard and adaptive median filtering algorithms, and others.
The peak signal-to-noise ratios of the filtered image using the various filtering techniques are computed quantitatively, to show the effectiveness and efficiency of the method of this thesis. The peak signal-to-noise ratio has been used to compare the performance of the proposed median filtering algorithm with other digital median filtering algorithms.
For example, the improved median filtering algorithm, when applied to the cameraman image after adding 20% impulse noise resulted in a filtered image with a peak signal to noise ratio of 38.0638. Applied to the same corrupted image, the adaptive median filter, the standard median filter, the maximum filter and the minimum filter yielded a filtered image with a peak signal to noise ratio of 33.9901dB, 32.1349dB, 25.9170dB and 30.3778dB, respectively.|
|Description: ||A thesis submitted to the Department of Computer Engineering in partial fulfilment of the requirements of Master of Philosophy in Computer Engineering, 2011|
|Appears in Collections:||College of Engineering|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.