Stop Thinking, Just Do!

Sungsoo Kim's Blog

Space-Time Video Completion

tagsTags

19 March 2017


Article Source


Space-Time Video Completion

The paper “Space-Time Video Completion” is available here:

http://www.wisdom.weizmann.ac.il/~vision/VideoCompletion.html

Unofficial implementation:

https://www.patreon.com/TwoMinutePapers

Abstrct

We present a method for space-time completion of large space-time holes in video sequences of complex dynamic scenes. The missing portions are filled-in by sampling spatio-temporal patches from the available parts of the video, while enforcing global spatio-temporal consistency between all patches in and around the hole. This is obtained by posing the task of video completion and synthesis as a global optimization problem with a well-defined objective function.The consistent completion of static scene parts simultaneously with dynamic behaviors leads to realistic looking video sequences. Space-time video completion is useful for a variety of tasks, including, but not limited to:

  • Sophisticated video removal (of undesired static or dynamic objects) by completing the appropriate static or dynamic background information.
  • Correction of missing/corrupted video frames in old movies
  • Synthesis of new video frames to add a visual story, modify it, or generate a new one. Some examples of these are shown in the paper.
  • Video texture synthesis
  • Full field-of-view stabilized video.

An implementation inspired by this work for filling images is available here.

NEW: This method is now part of the Content Aware Fill feature in Adobe Photoshop CS5.

  • A journal preprint from TPAMI March 2007, can be found here (pdf, 30MB).
  • A shorter version from CVPR 2004 is here (pdf, 3.5MB).

comments powered by Disqus