Efficient automatic 3D segmentation of cell nuclei for high-content screening

BMC Bioinformatics

Table 5 Results of segmentation of the artificial 3D stack specimens from Fig. 3

Image id/type		Confluency	TP	FP	FN	Prec	Sens	GT	Alg Cells	F1-score	\(JI_{Avg}\)
								Cells	Count
								Count
#1	3D image	Low	23	2	0	0.92	1	23	25	0.95	0.79
#2	3D image	Low	86	2	26	0.97	0.76	112	88	0.86	0.79
\(S_1\)	3D stack	Low	22	1	1	0.95	0.95	23	23	0.95	0.79
			86	2	26	0.97	0.76	112	88	0.86	0.79
#11	3D image	Low	39	0	1	1	0.97	40	39	0.98	0.85
#12	3D image	Low	78	2	16	0.97	0.83	94	80	0.89	0.84
\(S_2\)	3D stack	Low	39	0	1	1	0.97	40	39	0.98	0.85
			76	5	18	0.94	0.80	94	81	0.86	0.83
#10	3D image	Low	77	1	8	0.98	0.90	85	78	0.94	0.85
#11	3D image	Low	39	0	1	1	0.97	40	39	0.98	0.85
#12	3D image	low	78	2	16	0.97	0.83	94	80	0.89	0.84
#13	3D image	Mod	88	3	6	0.96	0.93	94	91	0.93	0.84
\(S_3\)	3D stack	Low/	79	0	6	1	0.92	85	79	0.96	0.84
		Mod	37	1	3	0.97	0.92	40	38	0.95	0.86
			70	7	24	0.90	0.74	94	77	0.81	0.80
			54	3	40	0.94	0.63	94	57	0.76	0.69

ISSN: 1471-2105