Video Operations in the Gradient Domain

Iddo Drori    Tommer Leyvand     Shachar Fleishman    Daniel Cohen-Or    Hezy Yeshurun

Technical Report, May 2004

Fusion of image sequences is a fundamental operation in numerous video applications and usually consists of segmentation, matting and compositing. We present a unified framework for performing these operations on video in the gradient domain. Our approach consists of 3D graph cut computation followed by reconstruction of a new 3D vector field by solving the Poisson equation. We introduce new methods for fusing video smoothly and separating foreground elements from a video background for compositing by defining smooth and sharp transition constraints on a gradient video compositing equation. We demonstrate the applicability of smooth video transitions by fusing pairs for video mosaics, video folding, and video texture synthesis, and demonstrate the applicability of sharp video transitions by video segmentation, video trimap extraction and 3D compositing into a new sequence. Our results demonstrate that our method maintains coherence of the video matte and composite, and avoids temporal artifacts.

PDF (7.2MB)

Some results:


Folding of the video sequence on itself, removing
the center tree. (cut) is the graph-cut result, (gvc)
is the result of our gradient video composition

Temporal composition of two sequences
resulting in a seamless morph.

Video matting using iterative 3D graph-cut

Apple QuickTime video sequences: Vid1 - Video folding (17MB), Vid2 (4.3MB) Vid3 (4.2MB) - Video matting

Back to my homepage