Quantitative PET image reconstruction employing nested expectation-maximization deconvolution for motion compensation

•A 3D motion-compensated PET image reconstruction algorithm employing nested Richardson-Lucy deconvolution is proposed.•Algorithm inspired from resolution modeling reconstruction techniques after replacing PSF kernel with an intra-frame motion blurring kernel.•Clinical adoptable method for both static and dynamic PET sinogram frames without requiring gating or access to list-mode data.•Applicable for any combination of rigid and non-rigid transformations including brain, cardio-respiratory and irregular bulk body motion.•Recommended for simultaneous PET/MR or dynamic whole-body PET scan protocols.