Medical image denoising based on 2D discrete cosine transform via ant colony optimization

In medical imaging, researchers usually encounter with different types of noise; for eliminating this noise, different methods have been suggested in both spatial and frequency domains. In this paper, a new method is proposed for removing Gaussian noise from medical images using two dimensional discrete cosine transform (2DDCT) and ant colony optimization (ACO) algorithm. In this algorithm, we attempted to identify the important frequency coefficients with the use of ant colony optimization and to eliminate the effects of noise by removing high frequency parts. Our proposed algorithm have been tested for various densities of the Gaussian noise and the experimental results show performance improvement in terms of peak signal to noise ratio (PSNR) and structural similarity (SSIM). (C) 2017 Elsevier GmbH. All rights reserved.