A self-adaptive image segmentation method, belonging to the unsupervised field. A target domain image is converted, such that semantic information thereof is reserved, shallow layer information of the image is reconstructed by taking a source domain image as a feature, and image discrimination is then performed on the converted and reconstructed target domain image by using a segmentation model established by the source domain image, thereby realizing migration between data sets of different fields of a model on the premise of no new data annotation.
展开▼