Brain image denoising using dual-channel attentional residual network

In medical imaging, noise interference reduces brain image quality and interpretation. Conventional noise reduction techniques, while reducing noise, usually lose image details and require specific filters, increasing complexity and limiting use. In this paper, based on deep convolutional neural networks, we design a dual-channel attentional residual network for brain images denoising model using the combination of channel attentional mechanism and spatial attentional mechanism, by introducing attentional mechanism in each of the four residual blocks and optimizing the network parameters using multiple loss functions. The model effectively preserves image details while removing noise, improving the quality and usability of brain images. Experimental results show that the method mostly outperforms other methods in the three evaluation metrics. The results of this research have important implications for the diagnosis and treatment of brain diseases.