Image denoising using wave atom transform

The image usually has different kinds of noises which are not easily removed in process of receiving, coding and transmission. Wave atoms is one of the new geometric multiscale-multidirectional transform, after 1999 that is based on wavelet transform, whose structural elements involve the parameters of dimension and location, and orientation parameter more, which let wave atom transform has good orientation characteristic. For that reason, wave atom transform is better than wavelet and even curvelet transform. In the expression of image edge, such as geometry characteristic of curve and beeline, which has previously obtained excellent research results in image denoising. This paper introduced a new method for image de-noising by coefficient thresholding based on wave atom transform. Wave atom proposes a new representation have a higher performance than the other transform. Experiments results indicate that the improved wave atom transform has abroad future for eliminating the noise of images. It remains edges efficiently, and offers a better visual effect.