Dynamic cardiac MRI requires rapid image acquisition to freeze heart motion. Accelerated scanning undersamples the data, creating aliasing artifacts that must be resolved computationally. Standard methods — compressed sensing, neural networks — trade reconstruction time for image quality.
Note (2026-11-07): the abstract of arXiv:2603.21980 proposes an explicit representation approach with multiple trained Gaussian primitives extended to model anatomical motion, reports that the splats can be trained in 60 seconds with 0.5 ms per dynamic of inference and that high-quality cardiac MRI is obtained at R = 16, and says the authors show the Gaussians' properties directly encode physiological properties. It does not compare explicit with implicit representations or say the former outperform the latter, or that standard methods trade reconstruction time for quality; the through-claim is mine. I compared this essay with the abstract only.
The authors (arXiv:2603.21980) repurpose Gaussian splatting, originally developed for 3D scene rendering, as a cardiac MRI reconstruction method. Multiple Gaussian primitives are trained to represent tissue properties, with the splatting framework extended to model anatomical motion. The result: high-quality cardiac images from 16× undersampled data, with training completing in 60 seconds and inference in 0.5 milliseconds.
The unexpected finding: the properties of the learned Gaussians — their positions, widths, and intensities — directly encode physiological properties of the heart. The representation isn't just a compression trick; it maps to anatomy.
The through-claim: explicit representations outperform implicit ones when the underlying structure has natural decomposition into discrete objects. The heart is made of distinct tissue regions, and Gaussian splatting's representation — a set of localized, parameterized components — matches that structure. The model succeeds not by brute-force learning but by structural correspondence between the representation and the anatomy.