Skip to main content
. 2024 Feb 12;14:3522. doi: 10.1038/s41598-024-53528-9

Table 9.

Final validation metric value (VMV) obtained after the learning paradigm optimisation phase compared to the initial VMV obtained after the pre-processing and data augmentation phase.

Model Initial VMV Final VMV Difference
nnFormer 70.4 69.7 −0.7
nnUNet 69.7 70.9 +1.2
SegmentationNet 66.3 65.2 −1.1
Swin-UNETR 63.2 62.8 −0.4
TransBTS 60.4 66.4 +6.0
UNETR 43.8 45.8 +2.0
VT-UNet 61.1 58.5 −2.6
Average 62.1 ± 8.3 62.8 ± 7.9 +0.6