Skip to main content
. Author manuscript; available in PMC: 2018 Jul 3.
Published in final edited form as: Med Image Anal. 2017 Oct 5;43:98–111. doi: 10.1016/j.media.2017.10.002

Table 3.

Evaluation results of our segmentation model with n = 1,2,3 (The number of patches used to train FCNNs was 5000*5*20.).

Dataset n Methods Dice PPV Sensitivity



Complete Core Enhancing Complete Core Enhancing Complete Core Enhancing
Challenge 1 FCNNs 0.70 0.62 0.64 0.57 0.50 0.55 0.94 0.87 0.79
FCNN + CRF 0.83 0.78 0.69 0.85 0.79 0.63 0.83 0.79 0.79
FCNN + CRF + post-process 0.86 0.83 0.76 0.92 0.88 0.77 0.82 0.80 0.77
3 FCNNs 0.71 0.67 0.65 0.58 0.56 0.57 0.95 0.86 0.78
FCNN + CRF 0.83 0.77 0.69 0.83 0.74 0.63 0.85 0.82 0.79
FCNN + CRF + post-process 0.86 0.82 0.76 0.91 0.84 0.78 0.84 0.82 0.76
5 FCNNs 0.74 0.72 0.67 0.62 0.63 0.60 0.94 0.86 0.77
FCNN + CRF 0.85 0.80 0.70 0.87 0.80 0.63 0.84 0.81 0.80
FCNN + CRF + post-process 0.87 0.83 0.76 0.92 0.87 0.77 0.83 0.81 0.77
Leaderboard 1 FCNNs 0.64 0.54 0.49 0.50 0.45 0.42 0.96 0.76 0.71
FCNN + CRF 0.82 0.65 0.57 0.83 0.70 0.51 0.85 0.69 0.72
FCNN + CRF + post-process 0.85 0.72 0.61 0.88 0.76 0.59 0.84 0.76 0.70
3 FCNNs 0.67 0.57 0.52 0.54 0.50 0.45 0.96 0.75 0.70
FCNN + CRF 0.83 0.66 0.57 0.82 0.67 0.51 0.87 0.72 0.71
FCNN + CRF + post-process 0.86 0.72 0.62 0.88 0.74 0.60 0.86 0.79 0.69
5 FCNNs 0.70 0.61 0.54 0.58 0.57 0.49 0.96 0.74 0.67
FCNN + CRF 0.83 0.66 0.57 0.85 0.71 0.50 0.85 0.69 0.71
FCNN + CRF + post-process 0.86 0.73 0.62 0.89 0.76 0.64 0.84 0.78 0.68