Our Poisson integrated results without occlusion handling. We render the data and gradient terms for the closest left and right reference cameras for the novel view, and blend them according to the inverse distance of their cameras. In the end we solve a single Poisson problem for the combined term images.