Table 4.
ASHS vs. manual rater | Manual intra‐rater reliability | |||||
---|---|---|---|---|---|---|
Head | Body | Tail | Head | Body | Tail | |
CA1 | 0.777 (0.040) | 0.878 (0.029) | 0.805 (0.088) | 0.905 (0.017) | 0.911 (0.020) | 0.894 (0.031) |
CA2 | 0.482 (0.192) | 0.594 (0.157) | 0.820 (0.081) | 0.815 (0.113) | ||
CA3 | 0.537 (0.117) | 0.431 (0.168) | 0.881 (0.034) | 0.743 (0.103) | ||
CA | 0.770 (0.038) | 0.869 (0.032) | 0.904 (0.017) | 0.911 (0.022) | ||
DG | 0.788 (0.036) | 0.892 (0.030) | 0.796 (0.080) | 0.922 (0.016) | 0.929 (0.025) | 0.912 (0.029) |
SUB | 0.761 (0.051) | 0.747 (0.057) | 0.815 (0.034) | 0.846 (0.033) | ||
ERC | 0.831 (0.038) | 0.850 (0.036) | ||||
BA35 | 0.747 (0.073) | 0.818 (0.048) | ||||
BA36 | 0.827 (0.046) | 0.896 (0.020) | ||||
PRC | 0.850 (0.036) | 0.906 (0.017) | ||||
HIPP | 0.891 (0.018) | 0.920 (0.017) | 0.877 (0.071) | 0.941 (0.009) | 0.946 (0.013) | 0.938 (0.013) |
For each subject/side, the T2‐weighted slices marked as head, body, or tail in the manual segmentation are extracted and DSC is computed only within those slices. Average DSC is reported across all cross‐validation experiments.