Analysis of a deep learning-based method for generation of SPECT projections based on a large Monte Carlo simulated dataset

Table 6 Analysis of the simulations of the physical phantom measurements

Projections		Synthetic 64 U1		Synthetic 32 U4		Simulated noisy
Projections		SSIM	NRMSE (%)	SSIM	NRMSE (%)	SSIM	NRMSE (%)
Jaszczak	Simulated noise-free	0.998	1.94	0.998	2.20	0.973	7.70
Jaszczak	Simulated noisy	0.975	7.70	0.975	7.56	1	0
NEMA	Simulated noise-free	0.997	0.43	0.995	0.49	0.990	0.93
NEMA	Simulated noisy	0.987	0.94	0.985	1.05	1	0

SSIM and NRMSE values between synthetic projections generated based on simulated noisy projections and simulated noise-free and noisy projections, respectively