Multi-modal medical image registration method based on double constraints

The invention belongs to the technical field of medical image processing, and particularly relates to a multi-modal medical image registration method based on double constraints. Firstly, a brain magnetic resonance image data set is acquired and preprocessed; then constructing a registration network, and performing iterative training on the registration network by using the preprocessed brain magnetic resonance image data set; and finally, registering the multi-modal brain magnetic resonance image to be registered by using the trained registration network. Wherein in the processing process of the registration network, image boundary information is enhanced through a boundary enhancement network, the similarity between two images is effectively measured through a multi-modal joint loss function, and a registration deformation field is promoted to become smoother through folding point optimization, so that the accuracy and authenticity of registration are effectively improved. 本发明属于医学图像处理技术领域，具体涉及一种基于双约束的多模态医学图