REAL ORIENTED 2-D DUAL TREE WAVELET TRANSFORM WITH NON-LOCAL MEANS FILTER FOR IMAGE DENOISING

Abstract

A new image denoising method based on dual tree real oriented wavelet transform using Non-Local Means Filter (NLMF) is proposed. Real oriented 2-D Dual Tree Wavelet Transform (DTWT) of an image gives six sub bands which is non separable. Implementation of Real oriented 2-D DTWT requires only addition and subtraction of respective sub bands [10]. NLMF is applied to the low pass sub bands to estimate the uncorrupted intensity as an average weight of all pixels in the low pass sub band. Denoising performance is further improved by soft thresholding the sub band coefficients. Comparative evaluation shows that the denoising capability of the proposed method is better than other existing techniques in terms of Peak Signal to Noise Ratio (PSNR).