3D Shape-Weighted Level Set Method for Breast MRI 3D Tumor Segmentation

Table 2

Accuracy, specificity, and false alarm rate (%) of the different algorithms and noise densities in 3 × 3 mask blurred computer-simulated images.

Noise density		Methods
		ACM	LSM	SLSM	3D-LSM	3D-SLSM
5%	CCR	90.43	99.24	99.34	99.7	99.76
	SP	89.93	99.2	99.76	99.69	99.96
	FAR	10.13	0.8	0.7	0.31	0.26
10%	CCR	90.07	99.13	99.19	99.6	99.67
	SP	89.59	99.09	99.65	99.51	99.95
	FAR	10.52	0.92	0.85	0.42	0.35
15%	CCR	91.99	98.35	99.12	99.45	99.61
	SP	91.63	98.27	99.63	99.44	99.94
	FAR	8.48	1.75	0.93	0.58	0.42
20%	CCR	91.02	98.52	98.93	99.07	99.52
	SP	90.64	98.48	99.54	99.06	99.93
	FAR	9.51	1.56	1.13	0.98	0.51
25%	CCR	90.3	98.02	98.74	98.6	99.41
	SP	89.89	97.95	99.41	98.56	99.92
	FAR	10.27	2.1	1.33	1.49	0.62
30%	CCR	90.95	96.83	98.61	97.82	99.32
	SP	90.63	96.7	99.37	97.74	99.93
	FAR	9.58	3.35	1.47	2.31	0.72
35%	CCR	90.02	95.22	98.34	98.48	99.16
	SP	89.65	95.02	99.28	98.54	99.91
	FAR	10.57	5.06	1.76	1.61	0.89
40%	CCR	90.52	94.18	98	97.75	98.96
	SP	90.22	94.05	99.22	98.01	99.91
	FAR	10.04	6.17	2.1	2.38	1.11
45%	CCR	89.69	95.9	97.73	96.59	98.7
	SP	89.32	96.05	99.16	96.84	99.86
	FAR	10.92	4.35	2.4	3.61	1.38
50%	CCR	89.39	95.28	97.55	95.08	98.42
	SP	89.09	95.52	99.11	95.32	99.85
	FAR	11.23	5	2.6	5.22	1.67