Abstract: In this paper, a comparative analysis is performed on four different thresholding algorithms namely, Average filtering, Visu Shrink, Sure Shrink and Adaptive wavelet packet (WP) thresholding function for image de noising. These algorithms are implemented on different test images and for different amounts of noise intensities added to those images, and their performance are evaluated based on parameters like mean square error (MSE), peak signal-to-noise ratio (PSNR). On the basis of experimentally obtained result the best method is proposed. The proposed method applies multilevel WP decomposition to noisy images to obtain the optimal wavelet basis, using Shannon entropy. It selects an adaptive threshold value which is level and sub band dependent based on analyzing the statistical parameters of sub band coefficients. WP transform (WPT) along with optimal wavelet basis (OWB) for image decomposition is used. Next, a thresholding function is used to shrink small coefficients leading to calculate a modified version of dominant coefficients. The modification is done using optimal linear interpolation between each coefficient and the mean value of the corresponding sub band followed by reconstruction of de noised image from the corrected coefficients.

Keywords: Adaptive wavelet packet thresholding; Optimal wavelet basis (OWB) ; Shannon entropy; Wavelet packet transform(WPT),Discrete wavelet transform(DWT).