Table 2 The effects of the calcification surface meshing algorithms on the final tetrahedral mesh

From: Robust automated calcification meshing for personalized cardiovascular biomechanics

Mesh technique

Success (%) ↑

Dice ↑

HD 95% (mm) ↓

CD (mm) ↓

Merged nodes ↑

∣J∣↑

Bool: pymadcad

0.0

N/A

N/A

N/A

N/A

N/A

Bool: OpenSCAD

0.0

N/A

N/A

N/A

N/A

N/A

Bool: vtk

0.0

N/A

N/A

N/A

N/A

N/A

Bool: blender

3.7

N/A

N/A

N/A

N/A

N/A

Bool: cgal

7.4

0.753 ± 0.140

1.366 ± 0.518

0.316 ± 0.126

333 ± 360

0.332 ± 0.004

Bool: igl

70.4

0.811 ± 0.061

1.507 ± 0.446

0.364 ± 0.074

2480 ± 1552

0.449 ± 0.043

Bool: corefinement

74.1

0.811 ± 0.059

1.503 ± 0.435

0.365 ± 0.073

2553 ± 1545

0.453 ± 0.043

Bool: cork

75.2

0.812 ± 0.056

1.480 ± 0.374

0.367 ± 0.067

2827 ± 1762

0.456 ± 0.039

DeepCarve

96.3

0.763 ± 0.060

2.405 ± 2.778

0.559 ± 0.251

630 ± 231

0.606 ± 0.020†

DMTet (no opt)

100.0

0.830 ± 0.045

1.000 ± 0.000

0.247 ± 0.020

2908 ± 1777

0.549 ± 0.013

DMTet (no opt) → ConstrainedRemesh

100.0

0.838 ± 0.043

1.000 ± 0.000

0.245 ± 0.021

2898 ± 1773

0.585 ± 0.011

DMTetOpt

100.0

0.849 ± 0.034

1.000 ± 0.000

0.238 ± 0.021

2907 ± 1777

0.584 ± 0.007

DMTetOpt → ConstrainedRemesh (C-MAC)

100.0

0.853 ± 0.033*†

1.000 ± 0.000*†

0.239 ± 0.022*†

2898 ± 1774†

0.593 ± 0.010*

  1. "Bool” refers to the mesh boolean difference operation between the original calcification surface and the aorta + aortic valve surface. All metrics were obtained using 10 repeated trials to account for the occasional failure of TetGen. We used the post-processed ground truth segmentation as the initial segmentation. Spatial accuracy metrics were measured against the initial segmentation, so we are strictly measuring the error induced by the meshing steps. Merged nodes indicate the number of the final calcification mesh nodes within 1e−3 mm from the CV tissue mesh nodes. All metrics are calculated using samples with non-empty targets only. All metrics are mean ± stdev across patients. Bold indicates best in each column.
  2. *p < 0.05 between C-MAC and “Bool: corefinement”; †p < 0.05 between C-MAC and DeepCarve.
Back to article page