A technique includes determining a depth value for each of a plurality of pixels of a frame, down-sampling the depth values of a tile of the frame to obtain a plurality of down-sampled depth values, the frame including one or more tiles, determining a change in a head pose, determining, from the plurality of down-sampled depth values, a down-sampled depth value for a vertex, determining an adjusted position for the vertex based on the change in head pose and the down-sampled depth value for the vertex, performing, based on at least the adjusted position for the vertex, a depth-adjusted time-warping of the frame to obtain a depth-adjusted time-warped frame, and triggering display of the depth-adjusted time-warped frame.