Skip to main content
. Author manuscript; available in PMC: 2022 Nov 14.
Published in final edited form as: Biomed Image Regist (2022). 2022 Jul 9;13386:103–115. doi: 10.1007/978-3-031-11203-4_12

Fig. 3.

Fig. 3.

Test Dice (left) and endpoint error (right) statistics on 10 structures across 500 T1w brain images. Regardless of the choice of the training loss function, the SuperWarp produces better Dice and endpoint error than the baseline (similar to VoxelMorph). Note that Ours–MSE does not need or use segmentation information.