TABLE 2.
Name | Modalities used | Commentaries | Hyper-parameters | PM threshold |
BIANCA | FLAIR | Preprocessing included except for WM mask computing | Lesion/non-lesion points: 2.000/10.000; WM mask threshold: 0.30 | 0.60 |
LST LGA | T1w, FLAIR | Preprocessing included | κ = 0.10 | 0.35 |
LST LPA | FLAIR | Preprocessing included | None | 0.10 |
LST LPA | T1w, FLAIR | Preprocessing included | None | 0.15 |
pgs | T1w, FLAIR | Training dataset from MICCAI challenge coregistration T1w to FLAIR | None | 0.15 |
sysu_media (default) | FLAIR | Training dataset from MICCAI challenge | None | 0.65 |
sysu_media (retrained) | FLAIR | Preprocessing: brain extraction, SANLM training dataset with 20 participants from DELCODE | None | 0.45 |
PM, probability map; WM, white matter; T1w, T1-weighted image; FLAIR, fluid-attenuated inversion recovery image; SANLM, spatially adaptive non-local means noise correction; MICCAI, the medical image computing and computer assisted intervention; LST LGA/-LPA, lesion segmentation toolbox for SPM using lesion growth algorithm/lesion prediction algorithms; BIANCA, brain intensity abnormality classification algorithm.