However, to implement this accurately, one would have to combine pixels from multiple frames in order to "see" these houses from different angles and combine those into a current frame. I am not sure how technically feasible this is, especially since those bridges might obscure visibility of some houses from certain angles.
Rotation that I am referring to is the https://en.wikipedia.org/wiki/Terrell_rotation which comes from a paper "Invisibility of the Lorentz Contraction" by James Terrell.
However, to implement this accurately, one would have to combine pixels from multiple frames in order to "see" these houses from different angles and combine those into a current frame. I am not sure how technically feasible this is, especially since those bridges might obscure visibility of some houses from certain angles.