In this paper we present a method to obtain a depth map from a single image of a scene by exploiting both image content and user interaction. Assuming that regions with low gradients will have similar depth values, we ...