Video inpainting with short-term windows: application to object removal and error concealment

Mounira Ebdelli 1        Olivier Le Meur 2        Christine Guillemot 1

1INRIA/IRISA, France
2univsersity of Rennes 1/IRISA

IEEE Transactions on Image Processing (2015)

[paper]

[soft. and datasets]

Home page O. Le Meur

Abstract

In this paper, we propose a new method for inpainting missing areas in static or free-moving camera videos. The method can be used for object removal, error concealment and background reconstruction applications. To limit the calculation time per frame, a frame is inpainted by considering a small number of neighboring pictures which are grouped into a group of pictures (GOP). More specifically, to inpaint a given target frame, the method starts by aligning all the frames of the GOP. This is achieved by a region-based homography computation method which allows to strengthen the spatial consistency of aligned frames. Then, from the stack of aligned frames, the inpainting process is performed by globally minimizing an energy function considering both space and time coherency terms. This energy function is efficient enough to provide high quality results even when the number of pictures in the GoP is rather small, e.g. 20 neighboring frames. This drastically reduces the algorithm complexity and makes the approach well suited for near real-time video editing applications as well as for loss concealment applications. Experiments with several challenging video sequences show that the proposed method provides more visually pleasing results than the most recent and efficient state-of-the-art approaches.


Object removal I :Comparison with Granados  Object removal II   Loss concealment   Background estimation   Failure cases   More results (Iterative inpainting, rigid transform)

Object Removal: comparison with Granados et al.(for a better viewing, display the video in fullscreen mode)

Left: Granados; Right: Our result
Sequence S5
Note that there is a color mismatch between the two results. It seems that the results provided by Granados et al. suffer from a compression problem or something like that. The original video can be watched below. For more details please visit the webpage http://people.mpi-inf.mpg.de/~granados/projects/vidbginp/index.html .


Left: Granados; Right: Our result
Sequence S4
Note that there is a color mismatch between the two results. It seems that the results provided by Granados et al. suffer from a compression problem or something like that. The original video can be watched below. For more details please visit the webpage http://people.mpi-inf.mpg.de/~granados/projects/vidbginp/index.html .


Left: Granados; Right: Our result
Sequence S6
Note that there is a color mismatch between the two results. It seems that the results provided by Granados et al. suffer from a compression problem or something like that. The original video can be watched below. For more details please visit the webpage http://people.mpi-inf.mpg.de/~granados/projects/vidbginp/index.html .

Object Removal (for a better viewing, display the video in fullscreen mode)

With Granados et al.'s video sequences

Top: Original; Bottom: Our result
Sequence S1


Sequence S3


Sequence S4


Sequence S5


Sequence S6


Sequence S7

With video sequences from Change Detection dataset

Badminton (Camera Jitter)


Boats (Dynamic Background)


Fountain02 (Dynamic Background)


Fall (Dynamic Background)


winterDriveWay (intermittent Object Motion)


peopleInShade (shadow)


Loss Concealment



Background estimation



Failure cases

Left: Granados; Right: Our result
Sequence S6
Note that there is a color mismatch between the two results. It seems that the results provided by Granados et al. suffer from a compression problem or something like that. The original video can be watched below. For more details please visit the webpage http://people.mpi-inf.mpg.de/~granados/projects/vidbginp/index.html .


More experimental results

Iterative inpainting

Left: One iteration; Right: Two iterations
Badminton (Camera Jitter)


Proposed approach vs an approach using a rigid transform

Left: proposed; Right: rigid transform
Boats (dynamic background)


Citation

Ebdelli, Mounira and Le Meur, Olivier and Guillemot, Christine (2015). Video inpainting with short-term windows: application to object removal and error concealment, submitted to IEEE Transactions on Image Processing, 2015.

Software and datasets

Related