Adaptive MRI image denoising using total-variation and local noise estimation

In this paper, we present an automated, adaptive image denoising method for removal of Rician noise from MRI images. The proposed method is based on the discretized total variation (TV) minimization model and the local noise estimation technique. The regularization parameter of the TV-based denoising method is adapted based on the standard deviation of noise in MRI image. The performance of the proposed method is evaluated using the brain MRI images corrupted by Rician noise with standard deviation ranging from 2 to 30. The quality of the denoised image is validated using both subjective visualization tests and objective quality metrics. The experimental results show that the proposed method achieves a significant improvement in the preservation of edges while simultaneously removing the Rician noise from a MR image. The adaptive TV filtering method is reasonably better than existing methods such as non-local filter, bilateral filter and multiscale linear minimum mean square-error estimation (LMMSE) approach.