Accept 180-degree and 360-degree spherical images as input

I think the feature might be already planned.

If you want to use them for creating geometry, the usual rules would still apply though.

Creating more than 6 images would achieve nothing at all since there would be no difference that could be used to calculate depth information.