Skip to main content
. 2020 Nov 20;14:610967. doi: 10.3389/fninf.2020.610967

Table 3.

Comparison of reviewed methods.

Methods Database Dim CNN style DSC PPV
Roy et al. (2018) ISBI 2015 3D Semantic-wise 0.524 0.866
Birenbaum and Greenspan (2016) ISBI 2015 3D Patch-wise 0.627 0.789
Valverde et al. (2019) ISBI 2015 3D Patch-wise 0.63 0.840
Aslani et al. (2019) ISBI 2015 2D Semantic-wise 0.61 0.899
Aslani et al. (2018) ISBI 2015 2D Semantic-wise 0.698 0.74
Zhang et al. (2019a) ISBI 2015 2.5D Semantic-wise 0.693 0.908
Havaei et al. (2016) MICCAI 2008 2D Patch-wise 0.832 N/A
Valverde et al. (2017) MICCAI 2008 3D Patch-wise 0.871 0.786
Brosch et al. (2016) MICCAI 2008 3D Semantic-wise 0.840 N/A
Valverde et al. (2016) MICCAI 2016 3D Patch-wise 0.541 N/A
McKinley et al. (2016) MICCAI 2016 3D Semantic-wise 0.591 N/A
Kazancli et al. (2018) Proprietary 3D Patch-wise 0.575 N/A
La Rosa et al. (2020) Proprietary 3D Semantic-wise 0.60 0.64
Brosch et al. (2015) Proprietary 3D Semantic-wise 0.355 0.414
Gabr et al. (2020) Proprietary 3D Semantic-wise 0.95 N/A
Coronado et al. (2020) Proprietary 3D Semantic-wise 0.77 N/A
Zhang et al. (2018) Proprietary 2D Semantic-wise 0.672 0.724
Aslani et al. (2020) Proprietary 3D Semantic-wise 0.50 0.519
Gessert et al. (2020a) Proprietary 4D Semantic-wise 0.64 N/A
Gessert et al. (2020b) Proprietary 3D Semantic-wise 0.656 N/A
Zhang et al. (2019b) Proprietary 2D Semantic-wise 0.660 N/A