I imagine an eyeball would be fairly similar but I think we can also view it in a similar way that I've been describing a 4D tyre.
The eye is largely symmetrical apart from the optic nerve and I don't think there is not much reason to suspect that it would be similar in 4D.
So again I think we could cut a 3D eye up into inifinte vertical cross-sections and just spin these 360deg through the forward axis to complete each 4D vertical cross-section.
Because the 4D 'human' sees things from their inside front face, rather than from our planer outsides, I imagine it is okay to just morph these cross-sections through any one of our directions and they will still represent the same thing from a 4Ders perspective maybe.
So if you took all these 4D cross sections and morphed the front backwards and smaller and morphed the back forwards but bigger and the rest just followed...
Now, I'm not sure about this because the middle would end up at the front and the back and the front and back would both be in the middle?
In terms of perspective my approach gives priority to the forward perspective being at front whereas I'm guessing the OPs perspective puts the forward dimension inside the camera.
I just use the approach of putting the front dimension in one direction and the side directions in the other two axis and having the vertical as cross-sections.
With my approach, which I think helps to directly connect our 3D objects with 4D objects a little easier, would it be possible to then convert them into a projection model?
Maybe it is a little difficult when curves are involved?