Abstract
In this paper, an Adaptive Generalized Gaussian Distribution (AGGD) oriented thresholding function for image de-noising is proposed. This technique utilizes a unique threshold function derived from the generalized Gaussian function obtained from the HH sub-band in the wavelet domain. Two-dimensional discrete wavelet transform is used to generate the decomposition. Having the threshold function formed by using the distribution of the high frequency wavelet HH coefficients makes the function data dependent, hence adaptive to the input image to be de-noised. Thresholding is performed in the high frequency sub-bands of the wavelet transform in the interval [-t, t], where t is calculated in terms of the standard deviation of the coefficients in the HH sub-band. After thresholding, inverse wavelet transform is applied to generate the final de-noised image. Experimental results show the superiority of the proposed technique over other alternative state-of-the-art methods in the literature.
Highlights
Noise can corrupt the image through acquisition or transmission processes
Proceeding step is setting a threshold value to see which coefficients are within the interval characterized by the threshold value and which coefficients are beyond this interval
A new technique for image de-noising utilizing a unique threshold function shaped by a process of using the GGD obtained from the HH sub-band in the wavelet domain after 4 levels of decomposition is proposed in this paper
Summary
Wavelet based image de-noising has become very popular among other noise removing techniques. Applying wavelet transform leads to two types of coefficients which can be divided into important and non-important coefficients, with the former should be kept due to having the most important characteristics of the image and the latter should be discarded. Noise suppression in wavelet domain requires a suitable threshold value to remove small noisy components of high frequency sub-bands and preserve larger coefficients of the same sub-bands. In this regards, an appropriate thresholding function and a defined threshold value are needed to suppress the additive noise and keep the noise-free data
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
