VIDEO COMPLETION/EDITING


Video completion is of great importance to many applications such as video repairing, video editing and movie post-production. We propose a novel technique to fill in missing background and moving foreground of a video captured by a static or moving camera. Different from previous work which are based on processing in the 3D data volume, we slice the volume along the motion manifold of the moving object, and therefore reduce the search space from 3D to 2D, while still preserving the spatial smoothness and temporal coherence. In addition to the computational efficiency, based on geometric video analysis, the proposed approach is also able to handle real videos under perspective distortion, as well as common camera motions, such as panning, translation and zooming. The experimental results demonstrate that our algorithm performs comparably to 3D search based methods, while extending the current state-of-the-art repairing techniques to videos with significant camera motion and projective effects, as well as lighting changes.

Some results are shown below by using our video completion method. In these video, some persons are removed, leaving the damaged background as well as some damaged moving foreground, e.g., occluded person. Therefore, the missing pixels in the background and foreground should be restored.

  • Video completion under camera motion:
    In this video, the camera is spanning, and the closer person is removed. Pixels in the background and the occluded person are recovered. In some frames of the original video, the person is mostly occluded, e.g., frame 29, however is recovered correctly in our result. The motion cue in the video is efficiently used to predict the motion of the occluded person. Click on images to enlarge.

    Frame 1 Frame 20 Frame 23 Frame 26
    Frame 29 Frame 32 Frame 35 Frame 84

    Download the result video HERE.

  • Video completion with perspective distortion:
    When the foreground is moving from near to far, there is perspective distortion in the volume of the foreground. Therefore, affine rectification and metric rectification should be applied to the volume before the video completion. After the foreground volume is completed, it is post-warpped back and combined with the background. The blur effect in the result comes from the rectification and post-warpping. Click on images to enlarge.

    Frame 1 Frame 42 Frame 44 Frame 46
    Frame 48 Frame 50 Frame 52 Frame 90

    Download the result video HERE.

  • Video completion with periodic illumination change:
    Our method can also handle the video with periodic illumination changes. In the video shown below, the global illumination is changing due to whirling of the fan. The the walking person is removed, and the occluded wall and ceiling, which are under changing illumination, should be completed. Click on images to enlarge.

    Frame 0 Frame 10 Frame 20 Frame 30
    Frame 40 Frame 50 Frame 60 Frame 70

    Download the result video HERE.