Fig. 1. Overall framework diagram

Fig. 2. MC-Net registration model diagram

Fig. 3. Schematic diagram of multi-scale constraint loss function

Fig. 4. Preprocessing diagrams

Fig. 5. Registration results of OASIS and LPBA40 datasets. (a)(e) Moving images to be registered; (b)(f) fixed reference images;

Fig. 6. Registration results of Abdomen CT-CT dataset. (a) Moving images to be registered; (b) fixed reference images; (c) registered images; (d) label maps of the moving images to be registered; (e) label maps of the fixed reference images; (f) label maps of the registered images

Fig. 7. Multi scale visualization of deformation field. (a) Moving images to be registered; (b) deformation fields (voxel: 160×192×160); (c) deformation fields (voxel: 80×96×80); (d) deformation fields (voxel: 40×48×40); (e) fixed reference images

Fig. 8. Comparison of parameter quantity before and after multi-core fusion

Fig. 9. Registration results of different methods on LPBA40 dataset. (a) Moving images to be registered; (b) fixed reference images; (c) registration results of ANTs; (d) registration results of VoxelMorph; (e) registration results of CycleMorph; (f) registration results of TransMorph; (g) registration results of MC-Net

Fig. 10. Registration results of different methods on the Abdomen CT-CT dataset. (a) Moving images to be registered; (b) fixed reference images; (c) registration results of ANTs; (d) registration results of Elastix; (e) registration results of VoxelMorph; (f) registration results of TransMorph; (g) registration results of MC-Net

Fig. 11. Ablation experimental results on the LPBA40 dataset. (a) Moving images to be registered; (b) registration results of U-Net; (c) registration results of U-Net+MK; (d) registration results of U-Net+MK+CBAM; (e) fixed reference images