An effective example-based denoising method for CT images using Markov random field

We propose in this paper a novel example-based method for Gaussian denoising of CT images. In the proposed method, denoising is performed with the help of a set of example CT images. We construct, from the example images, a database consisting of high and low-frequency patch pairs and then use the Markov random field to denoise. The proposed denoising method can restore the high-frequency band that is often lost by the traditional noise-filters. Moreover, it is very effective for images corrupted by heavy noise. Experimental results also show that the proposed method outperforms other state-of-the-art denoising methods both in the objective and subjective evaluations.