Video matting, or layer extraction, is a classic inverse problem in computer vision that involves the extraction of foreground objects, and the alpha mattes that describe their op...
Abstract--Modeling background and segmenting moving objects are significant techniques for video surveillance and other video processing applications. Most existing methods of mode...
In a known environment, objects may be tracked in multiple views using a set of background models. Stereo-based models can be illumination-invariant, but often have undefined valu...
Trevor Darrell, David Demirdjian, Neal Checka, Ped...
Abstract. Given an image, digital matting consists in extracting a foreground element from the background. Standard methods are initialized with a trimap, a partition of the image ...
This paper presents a method for deblurring an image consisting of two layers (a foreground layer and a background layer) which have suffered different, unknown blurs. This is a s...