A Comparitive Analysis of Thresholding Techniques Used in Image Denoising through Wavelets

Abstract

Historically, the field of image processing grew from electrical engineering as an extension of the signal processing branch. The massive amount of data required for images is a primary reason for the development of many sub areas within the field of computer imaging such as image segmentation and compression. Whatever may be the way of transmission, the data tends to get noisy and thereby the further processing does not lead to good results. Hence, it is very essential to keep the data close to originality. The prime focus of this thesis is related to the pre processing of an image. The pre processing being worked upon is the de noising of images. In order to achieve this in terms of the concerned work, wavelet transforms have been applied: Discrete wavelet transform and Un decimated Discrete wavelet transform. In this thesis, a new thresholding technique has been presented alongwith the standard thresholding techniques like soft and hard thresholding. And a comparative analysis of different combinations of the suggested threshold values and thresholding techniques has been carried out very efficiently. A new constraint, of either thresholding the low pass components or keeping them as such before applying the inverse DWT and UDWT, has also been added. This has been done in order to find more possible combinations that can lead to the best denoising technique. MATLAB codes have been developed for all the possible combinations, separately.