Skip to main content
. 2025 Jun 14;20(7):1551–1560. doi: 10.1007/s11548-025-03405-1

Table 3.

Ablation studies of knowledge distillation (KD), data generation (DG), and contrastive learning (CL) under DeepLabV3+ and MedSAM frameworks

Method DeepLabV3+ Framework MedSAM Framework
1–7 8–11 1–11 1–7 8–11 1–11
KD 62.33 45.16 56.09 61.86 35.93 52.43
KD+DG 66.94 42.80 58.16 63.25 36.66 53.58
KD+DG+CL 68.79 47.34 60.99 64.07 40.30 55.43