CT Denoising by Multi-feature Concat Residual Network with Cross-domain Attention Blcok

Computed Tomography (CT) is widely used in medicine, which has an irreplaceable role compared with other medical imaging methods because of its fast imaging speed, low cost and good imaging effect on bone and lung. But X-rays are harmful to the human body. In order to reduce the harm caused by the process of obtaining CT, low-dose CT is gaining popularity in recent years. Guided by deep learning, low-does CT denoising is successful using artificial neural network. This paper will use the convolutional neural network (CNN), combined the attention block and perceptual loss, to achieve excellent low-does CT denoising performance while preserving more details. Experimental results show that our method achieves good results at different noise levels.