Spatiotemporal redundancy removal in immersive video coding