Non-Decimated Wavelet Shrinkage Algorithm for Image Denoising Based on Inter-Scale Correlation

Abstract

This paper presents an image denoising scheme based on the correlation of the wavelet coefficients. It is well settled that significant features in images evolve with high magnitude across wavelet scales, while noise decays rapidly. Multiplying the adjacent wavelet scales sharpens the edges structures while weakening noise. This property is exploited by applying threshold to the scale correlation to identify the important features. Non-decimated wavelet transform is used here. Experiments shows that proposed method gives better results compared to other related works.