Table 3 Performance comparison of skull stripping for different stage MRIs of UPMC data.
From: A general skull stripping of multiparametric brain MRIs using 3D convolutional neural network
Type of MRI | # of cases | Dice | Precision | Recall | FPR | FNR | HD95 (mm) | p-value (t-test) |
|---|---|---|---|---|---|---|---|---|
Healthy MRIs (Healthy) | 57 | 0.9851 ± 0.0139 | 0.9963 ± 0.0070 | 0.9745 ± 0.0260 | 0.0007 ± 0.0013 | 0.0255 ± 0.026 | 2.4399 ± 1.568 | 0.3869 (vs Post-op) |
Pre-operative MRIs (Pre-op) | 57 | 0.9906 ± 0.0068 | 0.9910 ± 0.0111 | 0.9904 ± 0.0097 | 0.0019 ± 0.0027 | 0.0096 ± 0.0097 | 2.1655 ± 1.5278 | 0.0204 (vs Healthy) |
Post-operative MRIs (Post-op) | 63 | 0.9894 ± 0.0213 | 0.9942 ± 0.0080 | 0.9852 ± 0.0358 | 0.0011 ± 0.0016 | 0.0148 ± 0.0358 | 2.1751 ± 3.6873 | 0.2301 (vs Pre-op) |
