How would we represent an image with a different type of camera? Like a wide angle or fisheye lenses, would the math/computation be much more involved or would it take about the same processing power?

I think it would require some combination of perspective projection and non-linear operations