We learn a deformable registration model for T1-weighted MR images by considering multiple image characteristics via a hybrid loss. Our method registers the OASIS dataset with high accuracy while preserving deformation smoothness.
翻译:我们通过考虑通过混合损耗的多重图像特征来学习T1加权MR图像的变形注册模式。 我们的方法非常精确地记录了OASIS数据集,同时保持了变形平稳。