The concurrent encoding of viewpoint-invariant and viewpoint-dependent information in visual object recognition