Skip to main content
. 2019 Dec 28;25:102149. doi: 10.1016/j.nicl.2019.102149

Table 1.

Lesion Detection Results: Comparison between the different models evaluated. The results represent the mean detection TPF, FPF, DSCd, mean segmentation DSCs, and the mean runtime in minutes when analyzing the 36 MS patients using a leave-one-out cross-validation scheme. The automatic segmentation masks were obtained by thresholding the probability maps at 0.5 (using argmax), and all automatic lesions with a size lower than three voxels were removed.

Method TPF FPF DSCd DSCs Runtime (in minutes)
SimLearnedDFs 83.09  ±  21.06 9.36  ±  16.97 0.83  ±  0.16 0.55  ±  0.18 8.70  ±  0.09
SepLearnedDFs 57.77  ±  34.34 13.67  ±  21.99 0.60  ±  0.31 0.39  ±  0.22 9.08  ±  0.06
DemonsDFs 62.06  ±  32.74 11.98  ±  23.09 0.67  ±  0.29 0.42  ±  0.24 18.10  ±  0.05
NDFs 53.99  ±  38.01 17.20  ±  26.96 0.55  ±  0.35 0.37  ±  0.28 7.58  ±  0.09
Sweeney et al. (2013) 59.82  ±  37.59 33.59  ±  33.52 0.57  ±  0.33 0.44  ±  0.26 8.36  ±  0.01
Cabezas et al. (2016) 70.93  ±  34.48 17.80  ±  27.96 0.68  ±  0.33 0.52  ±  0.24 18.36  ±  0.02
Salem et al. (2018) 80.0  ±  27.77 21.87  ±  26.26 0.76  ±  0.25 0.55  ±  0.22 18.55  ±  0.02
Schmidt et al. (2019) 68.66  ±  35.26 31.89  ±  36.10 0.62  ±  0.34 0.40  ±  0.25 7.58  ±  0.03