Table 2 Quantitative comparison of cartilage thickness estimation across methods

From: CartiSurface: implicit surface reconstruction for anatomically-aware cartilage thickness mapping in knee MRI

Method

MAE (mm) ↓

Dice ↑

Hausdorff (mm) ↓

Smoothness ↓

U-Net + Morph9

0.52 (0.08)

0.83 (0.03)

3.41 (0.62)

0.218 (0.041)

SPHARM-PDM15

0.47 (0.06)

0.85 (0.04)

3.18 (0.58)

0.190 (0.037)

SurfaceFlow39

0.39 (0.05)

0.88 (0.02)

2.95 (0.43)

0.162 (0.031)

SegResNet38

0.45 (0.07)

0.86 (0.03)

3.02 (0.49)

0.205 (0.036)

HD-BET45

0.48 (0.07)

0.84 (0.04)

3.20 (0.51)

0.212 (0.038)

TopoShapeNet30

0.37 (0.05)

0.87 (0.03)

2.71 (0.40)

0.139 (0.028)

ImplicitRecon40

0.34 (0.05)

0.89 (0.02)

2.42 (0.35)

0.118 (0.023)

CartiSurface (Ours)

0.28 (0.04)*

0.91 (0.02)*

2.13 (0.29)*

0.092 (0.018)*

  1. All values reported as mean (standard deviation). Best results in bold.
  2. *Denotes statistically significant improvement over the second-best (paired t-test, p < 0.01).
Back to article page