It is well known that noise in Magnetic Resonance Image has a Rician distribution. Unlike additive Gaussian noise, Rician noise is signal dependent, and separating signal from noise is a difficult task. In this paper, a denoising technique is used in order to remove Rician noise from MRI using Waveatom shrinkage. De-noising by any shrinkage technique is highly sensitive to the threshold selection. Here to estimate the noise variance, histogram based technique is used and to calculate the shrinkage threshold a new technique is proposed. This method is applied to both simulated images and real images. Wave atom transform has been applied for different noise levels. This has been done in order to find more accurate results. A comparative analysis of wave atom and wavelet is also performed.