Table 3 Number of Clicks (NoC) required to achieve specific thresholds for Dice Similarity Coefficient (DSC) and Hausdorff Distance at the 95th percentile (HD95) on the MDA and HECKTOR 2021 datasets.
From: Interactive 3D segmentation for primary gross tumor volume in oropharyngeal cancer
Method | 0.75 DSC | 0.85 DSC | 5.0 mm HD95 | 2.5 mm HD95 | ||||
|---|---|---|---|---|---|---|---|---|
NoC (cnt) | PoF (%) | NoC (cnt) | PoF (%) | NoC (cnt) | PoF (%) | NoC (cnt) | PoF (%) | |
(a) Number of Clicks (NoC) required to achieve specific thresholds on the MDA dataset (\(N = 67\)), using an ensemble of models trained via 5-fold cross-validation on the HECKTOR 2021 dataset. | ||||||||
2S-ICR (ours) | \(\mathbf {1.81 \pm 0.12}\) | \(\mathbf {0.00 \pm 0.00}\) | \(\mathbf {5.97 \pm 0.25}\) | \(\mathbf {15.92 \pm 1.41}\) | \(\mathbf {1.00 \pm 0.00}\) | \(\mathbf {0.00 \pm 0.00}\) | \(\mathbf {6.50 \pm 0.41}\) | \(\mathbf {17.41 \pm 1.86}\) |
\(2.50 \pm 0.06\) | \(1.49 \pm 0.00\) | \(7.40 \pm 0.29\) | \(20.90 \pm 2.11\) | \(2.00 \pm 0.00\) | \(0.00 \pm 0.00\) | \(8.00 \pm 0.00\) | \(21.89 \pm 1.41\) | |
DeepEdit-2513 | \(2.67 \pm 0.12\) | \(1.00 \pm 0.70\) | \(7.50 \pm 0.08\) | \(26.37 \pm 0.70\) | \(1.33 \pm 0.47\) | \(0.00 \pm 0.00\) | \(8.00 \pm 0.41\) | \(28.86 \pm 1.86\) |
DeepEdit-5013 | \(3.46 \pm 0.09\) | \(2.49 \pm 0.70\) | \(8.46 \pm 0.56\) | \(15.92 \pm 1.41\) | \(1.00 \pm 0.00\) | \(3.48 \pm 0.70\) | \(10.17 \pm 0.62\) | \(34.83 \pm 0.70\) |
(b) Number of Clicks (NoC) required to achieve specific thresholds on the HECKTOR 2021 dataset (\(N = 224\)), trained using 5-fold cross-validation. | ||||||||
2S-ICR (ours) | \(\mathbf {1.34 \pm 0.46}\) | \(1.64 \pm 1.72\) | \(\mathbf {3.96 \pm 0.97}\) | \(\mathbf {16.38 \pm 5.45}\) | \(0.07 \pm 0.25\) | \(\mathbf {0.30 \pm 0.77}\) | \(\mathbf {0.97 \pm 0.64}\) | \(\mathbf {6.27 \pm 3.11}\) |
\(2.05 \pm 0.41\) | \(\mathbf {0.45 \pm 0.90}\) | \(5.02 \pm 0.51\) | \(19.52 \pm 4.17\) | \(0.40 \pm 0.49\) | \(0.30 \pm 0.76\) | \(1.90 \pm 1.17\) | \(8.80 \pm 4.42\) | |
DeepEdit-2513 | \(2.41 \pm 1.16\) | \(1.05 \pm 1.12\) | \(4.97 \pm 1.13\) | \(25.13 \pm 13.32\) | \(0.33 \pm 0.47\) | \(0.45 \pm 0.91\) | \(4.20 \pm 3.86\) | \(10.01 \pm 5.90\) |
DeepEdit-5013 | \(2.44 \pm 0.65\) | \(1.20 \pm 1.80\) | \(4.54 \pm 0.91\) | \(25.00 \pm 7.08\) | \(\mathbf {0.00 \pm 0.00}\) | \(1.49 \pm 1.78\) | \(2.70 \pm 1.93\) | \(11.33 \pm 4.90\) |
