Improvement Of Quantized Adaptive Switching Median Filter For Impulse Noise Reduction In Grayscale Digital Image

Abdalameer, Ahmed Khaldoon (2017) Improvement Of Quantized Adaptive Switching Median Filter For Impulse Noise Reduction In Grayscale Digital Image. Masters thesis, Universiti Sains Malaysia.

[img]
Preview
PDF - Submitted Version
Download (1151Kb) | Preview

    Abstract

    In this dissertation, an improvement to Quantized Adaptive Switching Median filter (QSAM) has been done, to make it more efficient in reducing high density fixedvalued impulse noise from grayscale digital images. QSAM uses the switching approach, where it has noise detection and noise cancellation blocks. This approach minimizes unwanted changes from the filtering process. QSAM also uses adaptive approach, where the filter size is adaptable to the local noise content. QSAM has two main stages. In the first stage, the image is filtered using the filtering window with quantized size. In the second stage, the image is filtered using adaptive window size. Improvement to QSAM has been carried out by replacing the formula used to restore the corrupted pixel. Instead of using the local median value, this proposed method uses the average of the local mean and local median values. Experimental results using three standard grayscale images of size 512 512 pixels show that the proposed method has the ability to restore the corrupted images even up to 95% of corruption. As compared to other thirteen median filters, the proposed method had the lowest Mean Square Error (MSE) and produce outputs with the best visual appearance.

    Item Type: Thesis (Masters)
    Subjects: T Technology > TK Electrical Engineering. Electronics. Nuclear Engineering > TK1-9971 Electrical engineering. Electronics. Nuclear engineering
    Divisions: Kampus Kejuruteraan (Engineering Campus) > Pusat Pengajian Kejuruteraaan Elektrik & Elektronik (School of Electrical & Electronic Engineering) > Thesis
    Depositing User: Mr Mohd Fadli Abd Rahman
    Date Deposited: 08 Nov 2017 15:05
    Last Modified: 17 May 2018 11:09
    URI: http://eprints.usm.my/id/eprint/37339

    Actions (login required)

    View Item
    Share