Table 2.
Results of measurements of dice similarity coefficient (DSC) and normalized weighted spectral distance (nWSD) for Case 1 (manual registration) and Case 2 (automated registration)
N | Mean DSC ± CI 95% | SD | Min | Max | Mean nWSD ± CI 95% | SD | Min | Max | |
---|---|---|---|---|---|---|---|---|---|
Case 1 | |||||||||
2D warp | 103 | 0.9642 ± 0.0186 | 0.0955 | 0.0110 | 0.9878 | 0.0066 ± 0.0005 | 0.0024 | 0.001 | 0.01 |
3D warp | 1 | 0.9486 | – | – | – | 0.0045 | – | – | – |
Case 2 | |||||||||
2D warp | 262 | 0.9427 ± 0.0030 | 0.0291 | 0.7881 | 0.9780 | 0.0080 ± 0.00003 | 0.0002 | 0.008 | 0.009 |
3D warp | 1 | 0.9675 | – | – | – | 0.0054 | – | – | – |
Mean values with 95% confidence intervals (CI 95%), standard deviations (SD), and minimal and maximal values are also listed. N is the number of pairs of slices (2D) or pairs of volumes (3D) evaluated. t test demonstrated p < 0.05 for comparison between DSC and nWSD values for 2D warps of Cases 1 and 2