A novel image denoising method was proposed by incorporating the undecimated Curvelet wavelets into the ordinary curvelet transform.Curvelet coefficients were acquired by using orthogonal wavelet transform in radon domain.Traditional discrete wavelet transform is not shiftable

so it will lead to Gibbs phenomenon.The orthogonal wavelet was replaced with undecimated wavelet transform in the last step of the Curvelet transform.The shift invariant property of the undecimated wavelet and the high directional sensitivity of the curvelet transform made the new method a good choice for image denoising.The digital undecimated Curvelet transform was applied to denoise some standard images corrupted with additive Gaussian white noise.Experimental results showed that the proposed method performanced better than the ordinary curvelet image denoising

and wiener2 filter in terms of both peak signal-to-noise ratio and visual quality.In particular

the method preserved the shape edges better while removing white noise.