Super-resolution Algorithm Of Medical MRI Image Based On Generative Adversarial Network

With the development of modern medical diagnosis and treatment and computer technology,medical imaging became a powerful tool in clinical diagnosis and medical research.Medical MRI image with multi-directional and multi-parameter imaging mode had a high resolution of human brain structure and soft tissue,without ionizing radiation,and was the main means of pathological information and anatomical structure research.In practical application,due to the differences of imaging equipment and technology,as well as external interference and other factors,the acquired MRI image details were lost and the texture was not clear.Therefore,the study of super-resolution MRI image had important scientific significance for clinical diagnosis and human medical career.By analyzing the texture features of medical MRI images,according to the differences of feature information of images obtained by convolutional neural network under different receptive fields and the differences of complex texture information among tissues and organs of medical images,this dissertation adopted generative adversarial network with high fidelity image generation ability to construct the following two image super-resolution algorithms:（1）Super-resolution reconstruction algorithm for medical MRI image based on multi-scale residuals generative adversarial network.Firstly,the multi-scale residual group was used to improve the residual blocks in the network;the local residual feature aggregation module aggregated the residual groups together to realize the non-local use of residual features and reduce the loss of local features in the process of network transmission;then,the attention module aimed to enable the network to obtain channel and spatial feature in-formation with a higher degree of response to the key information,so as to improve the detail texture effect of reconstructed image;next,the gradient image of the low resolution image was transformed into the gradient image of the high resolution image to assist the reconstruction of the super-resolution image;finally,the restored gradient image was integrated into the super resolution branch to provide structural prior information for super resolution reconstruction,so as to clearly guide the generation of high quality super resolution image.（2）Super-resolution reconstruction algorithm for medical MRI image based on multi-receptive field generative adversarial network.Firstly,the multi-receptive field feature extraction block was used to obtain the global feature information of the images under different receptive fields.In order to avoid the loss of detailed texture due to too small or too large receptive fields,each group of features was divided into two groups,and one of which was used to feedback global feature information under different scales of receptive fields,and the other group was used to enrich the local detailed texture information of the next set of features;then,the multi-receptive field feature extraction block was used to construct feature fusion group,and spatial attention modules were added to each feature fusion group to adequately obtain the spatial feature information of the image,reduced the loss of shallow and local features in the network,and achieved a more realistic degree in the details of the image;secondly,the gradient map of the low-resolution image was converted into the gradient map of the high-resolution image to assist in the reconstruction of the super-resolution image;finally,the restored gradient image was integrated into the super-resolution branch to provide structural prior information for super-resolution reconstruction,which was helpful to generate high quality super-resolution images.Experiments show that the proposed methods of this dissertation are superior to other algorithms in both objective index and subjective vision.