> A lightweight neural network is then used to reconstruct a hole-free image including detail beyond splat resolution.

Seems like if you added a similar super resolution step to 3DGS you would get similar detail improvements?