Table 1:

One-shot segmentation benchmarking of performance (median IoU & HD95) and longitudinal consistency (ASPC). Further scores alongside means and std. dev. are provided in Supplemental Table 7. Few-shot and fully-supervised results are provided in Suppl. Tabs. 3 and 4, respectively.

Method	IBIS-subcort			IBIS-wmgm			OASIS3
	IoU ↑	HD95 ↓	ASPC↓	IoU↑	HD95↓	ASPC↓	IoU ↑	HD95 ↓	ASPC↓
GT	-	-	7.819	-	-	-	-	-	3.947
RandInitUnet.2D	0.707	4.485	7.127	0.510	3.274	11.506	0.687	2.545	10.189
RandInitUnet.3D	0.720	2.892	4.644	0.560	3.788	5.515	0.715	2.206	2.789
Context Restore [10]	0.711	4.403	7.831	0.444	8.273	29.235	0.717	3.323	5.577
LNE [43]	0.736	3.033	5.866	0.563	3.201	5.352	0.726	1.988	8.836
GLCL [9]	0.718	3.203	5.514	0.550	4.112	8.472	0.695	2.264	4.622
PCL [63]	0.713	3.270	5.610	0.562	4.974	10.648	0.707	2.327	4.850
PatchNCE [44]	0.743	1.266	5.780	0.607	4.344	3.782	0.738	2.275	3.114
Ours w/o ${ℒ}_{cs}$	0.754	1.145	5.483	0.614	3.291	2.462	0.739	1.940	2.729
Ours w/ ${ℒ}_{cs}$	0.757	1.178	4.475	0.676	3.237	4.155	0.737	2.094	2.754