Myocardial perfusion imaging (MPI) from single-photon emission computed tomography (SPECT), which is widely employed clinically, can provide critical diagnostic information on coronary artery disease Garcia 2001, Lin et al 2006). However, there are multiple factors such as imaging noise, limited spatial resolution, cardiac and respiratory motion blur that can degrade the quality of the reconstructed images, and thereby adversely affect the diagnostic accuracy in SPECT (Cooper et al 1992).To reduce the extent of cardiac motion blur, cardiac-gated acquisitions (Slomka et al 2004, Suzuki et al 2008, Song et al 2018a) are often used, wherein the acquired data are binned into a number of intervals within the cardiac cycle according to the electrocardiography (ECG) signal (Kortelainen et al 2015, Slomka andGermano 2017). While this can effectively reduce the cardiac motion blur within the individual intervals, it also lowers the data counts and thereby leads to increased reconstruction noise. Because of this, in the clinic, gated studies are primarily used for functional assessment of the left ventricular (LV) wall motion and thickening, whereas ungated studies are used for assessment of perfusion defects (Ramon et al 2017).