TABLE III.

A Comprehensive Summary of the Evaluation Results. For MBF Bias, * Indicates the Results’ Mean Is Significantly Different From Zero Based on Paired t-Test (5% Significance Level). Bold Font Indicates the Best Performer in Each Group

Error Type	Method	Square Motion(200)	Triangle Motion(200)	Spike Motion (200)
Mean Motion Error (mm)	NoMC	4.6±1.7	2.6±1.0	4.7±1.3
	MC-IR	6.4±2.2	6.4±2.2	6.4±2.1
	MC-CR	4.6±1.0	4.7±0.9	4.5±0.9
	MC-CNN	1.8±0.8	1.6±0.7	1.7±0.7
	MC-CNN-C	4.4±3.0	3.0±1.5	4.7±2.6
	DeepMC	1.0±0.6	0.9±0.5	0.8±0.4
Max Motion Error (mm)	NoMC	13.3±3.7	12.8±3.7	18.4±1.8
	MC-IR	46.7±19.7	46.4±19.9	46.2±19.2
	MC-CR	11.7±2.6	11.8±2.5	11.7±2.4
	MC-CNN	5.5±2.5	5.3±2.5	5.9±2.4
	MC-CNN-C	10.1±5.9	8.2±3.7	10.7±5.6
	DeepMC	3.6±2.0	3.2±1.9	3.6±1.9
MBF Bias (%)	NoMC	−9.8±33.0%*	−7.4±16.2%*	−4.0±17.3%*
	MC-IR	62.3±51.3%*	62.1±52.0%*	64.8±53.1%*
	MC-CR	6.0±7.4%*	5.9±7.3%*	6.1±6.9%*
	MC-CNN	1.2±5.9*	1.7±5.7*	2.2±5.6*
	MC-CNN-C	−2.3±10.4*	−2.8±12.5*	−3.0±10.6*
	DeepMC	−0.5±4.6%	−0.1±3.9%	0.4±3.6%
Fitting Residuals (Bq/mL×min2)	NoMC	1.43×10−3	9.54×10−4	2.16×10−3
	MC-IR	6.04×10−3	6.00×10−3	6.22×10−3
	MC-CR	1.81×10−3	1.77×10−3	1.81×10−3
	MC-CNN	3.33×10−3	2.92×10−3	3.11×10−3
	MC-CNN-C	7.47×10−4	7.82×10−4	7.93×10−4
	DeepMC	6.29×10−4	6.06×10−4	6.24×10−4
	Motion Free	6.20×10−4	6.20×10−4	6.20×10−4