An improved dual tree complex wavelet transform based image denoising using GCV thresholding

Noise suppression is an integral part of any image processing task Noise significantly degrades the image quality and hence makes it difficult for the observer to discriminate fine detail of the images especially in diagnostic examinations. Through decades of research, mass articles on image denoising have been proposed The effect of noise in the images can be reduced by using either spatial filtering or transform domain filtering. In transform domain wavelet method provide better denoising while preserving the details of images like edges. The Discrete Wavelet Transform (DWT) has some disadvantages that undetermined its application in image processing as lack of shift invariance and poor directional selectivity. In order to overcome these disadvantages Dual Tree Complex Wavelet Transform (DT-CWT) is used which provide perfect reconstruction over the traditional wavelet transform It employs 2 real DWTs; the first DWT gives the real part of the transform while second DWT gives the imaginary part. It is nearly shift invariant and directionally selective in two and higher dimensions with limited redundancy. The DTCWT outperforms the DWT for applications like image denoising and enhancement One of the advantages of the DTCWT is that it can be used to implement 2D wavelet transforms that are more selective with respect to orientation than is the 2D DWT. The 2D DTCWT produces twelve sub bands at each scale, each of which are strongly oriented at distinct angles. A Dual Tree Complex Wavelet transform based image denoising is proposed which uses generalized cross validation technique The denoising performance for different images using Discrete Wavelet Transform and Dual Tree Complex wavelet transform with different thresholding need to be evaluated Evaluation is carried out in terms of various parameters such as Peak Signal to Noise Ratio, mean Structural Similarity and Coefficient of Correlation.