Principal-channels for One-sided Object Cutout

We introduce principal-channels for cutting out objects from an

image by one-sided scribbles. We demonstrate that few scribbles, all

from within the object of interest, are sufficient to mark it out.

One-sided scribbles provide significantly less information than

two-sided ones. Thus, it is required to maximize the use of

image-information. Our approach is to first analyze the image with a

large filter bank and generate a high-dimensional feature space. We

then extract a set of principal-channels that discern one object

from another. We show that by applying an iterative graph-cut

optimization over the principal-channels, we can cut out the object

of interest.

In  Visual Computer

Publisher  Springer-Verlag New York, Inc.

Details

TypeArticle
Pages
Volume-
Number-
AddressSecaucus, NJ, USA
> Publications > Principal-channels for One-sided Object Cutout