DRDN: A Method for Low-Resolution Medical Image Denoising

In medical imaging, medical imaging inevitably generates noise due to uncontrollable factors such as principles, technical restrictions and biological environments. Even in some scenes, high-resolution images cannot be obtained, and only lowresolution images with much noise can be obtained. This results in unsatisfactory results of existing denoising algorithms. Therefore, improving the SNR of low-resolution medical images has become the key to solving the problem. In this paper, we propose the application of RDN to solve the problem of low-resolution denoising. The network is more efficient and more conducive to image detail restoration through training. Then, we evaluate our model and the classical CNN model on the public data set, and the experimental results show that the new DRDN can realize the denoising of low-resolution noise images at the noise level of 10. In the tests of 20 and 50, the PSNR is 40.21, 35.55, 31.08, and SSIM is 0.96, 0.93 and 0.83, respectively, which were superior to other classical models.