Translation Invariant Denoising Using Neighbouring Curvelet Coefficients

The denoising of a natural image corrupted by noise is a classical problem in image processing. Some curvelet denoising scheme have been introduced recently. However, they may discard some curvelet coefficients which may contain useful image information because of basing on uniform threshold and introduce many visual artifacts due to the pseudo-Gibbs phenomena. In this paper, we propose a new denoising scheme which is developed by combining a local adaptive shrinkage threshold based on the characteristic of neighbouring curvelet coefficients and cycle spinning technique. Experimental results show that the proposed approach outperforms uniform threshold method and local adaptive thresholding method without translation invariant in terms of the Peak Signal to Noise Ratio (PSNR) values and subjective image quality.