Cross-modal unsupervised domain adaptive medical image segmentation method

The invention discloses a cross-modal unsupervised domain adaptive medical image segmentation method, which is characterized in that image alignment and feature alignment between a source mode and a target mode are enhanced by adopting consistency regularization and uncertainty estimation, cross-modal unsupervised domain adaptive medical image segmentation is realized, and the method specifically comprises the following steps: 1) data preprocessing; 2) performing iterative training and parameter updating; 3) obtaining a segmentation model; 4) obtaining a segmentation result; and 5) evaluating a segmentation result. Compared with the prior art, the method has the advantages that the common features between the two modals can be learned under the condition that no target domain annotation is used, a good segmentation effect is achieved on the source modal image and the target modal image, the domain migration problem in the medical image is effectively solved, and the predicted segmentation image is more robust