Rician noise removal on MRI using wave atom transform with histogram based noise variance estimation

Wave atom transform is a new multi-resolution technique, which has the ability to adapt to arbitrary local directions of a pattern, and to sparsely represent anisotropic patterns aligned with the axes. In this paper, a de-noising technique is proposed to remove the rician noise from Magnetic Resonance Images using wave atom shrinkage. It is well known that the noise in magnetic resonance imaging has a Rician distribution. Unlike additive Gaussian noise, Rician noise is signal dependent, and separating signal from noise is a difficult task. 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 adapted. This method is applied to both simulated images and real images and is compared with wavelet shrinkage.