Abstract

We all know for a fact that various denoising algorithms have already been proposed. However, there is a keen interest of researchers to develop more effective techniques for image denoising. The proposed mechanisms have not been able to attain the desirable results. It is inevitable to stop the noise introduction while acquisition or transmission of the image. The majorly found noise in images is the Gaussian noise. The corruption of image with Gaussian noise is a classical problem till date. This paper can be called an extension to my previous work in which I have proposed a method for the removal of Gaussian noise using a hybrid of Neigh and Bayes thresholding techniques and Discrete Wavelet Transform. Here, we have extended it to more effective modified fast haar transform and applied a hybrid of post filters for better results. The main aim is to minimize the noise as much as possible. The results have been compared on various quality parameters such as PSNR, CNR, Standard Deviation and Entropy.

Authors and Affiliations