3D Liver Tumor Segmentation in CT Images Using Improved Fuzzy C-Means and Graph Cuts

Table 1

Quantitative evaluation of the proposed tumor segmentation method. Results are represented as mean ± standard deviation (VOE: volumetric overlap error; RVD: relative volume difference; ASD: average symmetric surface distance; RMSD: root mean square symmetric surface distance; MaxD: maximum symmetric surface distance; DICE: Dice similarity coefficient).

	VOE (%)	RVD (%)	ASD (mm)	RMSD (mm)	MaxD (mm)	DICE	Diameter (mm)
Small tumor	35.90 ± 7.06	0.38 ± 25.29	0.66 ± 0.23	0.99 ± 0.31	3.34 ± 1.27	0.78 ± 0.05	9.42 ± 2.11
Max	47.76	37.99	1.25	1.74	6.32	0.86	14.80
Min	24.11	−31.24	0.40	0.69	2.00	0.69	5.54
Large tumor	26.62 ± 7.11	−3.11 ± 11.05	0.74 ± 0.35	1.13 ± 0.54	4.57 ± 3.40	0.84 ± 0.05	16.95 ± 9.20
Max	46.50	18.25	2.40	3.27	20.78	0.95	45.38
Min	8.78	−18.07	0.40	0.68	1.73	0.70	7.61
Average	29.04 ± 8.16	−2.20 ± 15.88	0.72 ± 0.33	1.10 ± 0.49	4.25 ± 3.03	0.83 ± 0.06	14.99 ± 8.63