re-draw, then use the drawn geometry as the basis for a 3D model
manually sculpt a 3D model using the the photograph as a guide
use a Large Language Model (what folks are passing off as “A.I.”) to create a height/depth map or 3D model — while not a photo, here was a recent discussion of doing this sort of thing from a pixel image:
It might be that there are specialty models for this sort of thing, but that’s not something which I have looked into.