. Author manuscript; available in PMC: 2009 Dec 8.

Published in final edited form as: IEEE Trans Med Imaging. 2009 Jul 14;28(12):1955–1968. doi: 10.1109/TMI.2009.2026412

TABLE II.

Simulation results when random C-arm calibration and pose estimation errors are introduced. The results for each seed density are averaged over the source angle separations varying from 10° to 25° with a step size of 5° (total 800 simulations). The averaged results are rounded up to one digit after the decimal point which would be enough for clinically meaningful data analysis.

Num of seeds	Num of images	Number of seeds		Mean ± STD error (mm)
Num of seeds	Num of images	candidate	correctly detected	Mean ± STD error (mm)
54	3	57.6	53.7 (99.4%)	0.8 ± 0.5
54	4	54.3	54.0 (100.0%)	0.8 ± 0.3
60	3	65.4	59.7 (99.5%)	0.8 ± 0.4
60	4	60.5	59.9 (99.8%)	0.8 ± 0.2
72	3	80.5	71.3 (99.0%)	0.9 ± 0.6
72	4	72.9	71.9 (99.9%)	0.8 ± 0.3
84	3	96.4	82.7 (98.5%)	0.9 ± 0.5
84	4	85.5	83.7 (99.6%)	0.8 ± 0.3
96	3	108.3	93.9 (97.8%)	0.9 ± 0.6
96	4	100.8	95.3 (99.3%)	0.8 ± 0.3
112	3	128.8	108.3 (96.7%)	0.9 ± 0.7
112	4	118.9	110.6 (98.8%)	0.8 ± 0.4