Table 1.

Evaluation results of FCNNs with and without post-processing, FCNN + CRF with and without post-processing, and FCNN + 3D-CRF with and without post-processing. (The sizes of image patches used to train FCNNs were 33^*33^*3 and 65^*65^*3 respectively, n = 5, and the number of patches used to train FCNNs was 5000^*5^*20. FCNN + CRF is short for the integrated network of FCNNs and CRF-RNN).

Dataset	Methods	Dice			PPV			Sensitivity
		Complete	Core	Enhancing	Complete	Core	Enhancing	Complete	Core	Enhancing
Challenge	FCNNs	0.74	0.72	0.67	0.62	0.63	0.60	0.94	0.86	0.77
	FCNN + post-process	0.81	0.75	0.73	0.73	0.70	0.73	0.94	0.85	0.74
	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
	FCNN + 3D-CRF	0.85	0.80	0.73	0.84	0.78	0.69	0.88	0.83	0.80
	FCNN + 3D-CRF + post-process	0.87	0.83	0.78	0.89	0.85	0.78	0.86	0.83	0.79
Leaderboard	FCNNs	0.70	0.61	0.54	0.58	0.57	0.49	0.96	0.74	0.67
	FCNN + post-process	0.81	0.65	0.61	0.74	0.63	0.62	0.94	0.78	0.66
	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
	FCNN + 3D-CRF	0.84	0.65	0.61	0.81	0.71	0.57	0.90	0.71	0.71
	FCNN + 3D-CRF + post-process	0.87	0.71	0.63	0.88	0.74	0.63	0.88	0.79	0.70