Liver MRI image segmentation by fusing ResUnet and Pix2pix

Accurate segmentation of liver MRI images can help assist physicians in the early stages of the disease by qualitatively diagnosing the lesion area, and can also serve as an important assessment for postoperative testing. However, there are many difficulties in segmenting MRI images of the liver, such as low accuracy of traditional methods and poor adaptability to complex images. In order to solve these problems, this paper proposes a ResUnet-based Pix2pix segmentation model, which aims to reach the goal of accurate segmentation. Specifically, ResUnet uses the residual connection to strengthen the model and extract the features of MRI images, which are then used as the generator of the Pix2pix model and output the segmentation results, and finally optimized continuously under the supervision of the discriminator until an accurate segmentation of liver MRI images is obtained. The experimental results show that the proposed method can segment the liver cancer region more accurately, which provides a highly reliable basis for the clinical diagnosis and treatment of liver cancer.