We estimate dense, flicker-free, geometrically consistent depth from monocular video, for example hand-held cell phone video. - View it on GitHub