Curvelet and wavelet transform coupling for denoising images with white noise

In this paper the coupling method of time invariant wavelet and curvelet transforms was used for denoising of images with while noise. In the coupling method, the discrete curvelet transform is performed in three steps: splitting, tiling and ridgelet transforming. The ridgelet transform is designed to code linear singularities. It can not handle curved edges. The problem with hard thresholding is that by thresholding to zero all curvelet coefficients lower than given threshold causes wrinkles near discontinuities. This method does not zero out all curvelet coefficient that fall below the threshold value. By the experimentation of combining the attributes of both transformations, it can be seen that using this method to denoise images corrupted with while noise is better than using the wavelet and curvelet denoising methods individually.