ICCV
2007
IEEE
14 years 12 months ago
2007
IEEE
A variational approach for the background/foreground segmentation of multiple views of the same scene is pre
ICCV
2007
IEEE
14 years 12 months ago
2007
IEEE
Boosting has been widely applied in computer vision, especially after Viola and Jones's seminal work [23]. The marriage of rectangular features and integral-imageenabled fast...
ICCV
2007
IEEE
14 years 12 months ago
2007
IEEE
Correspondence establishment is a key step in statistical shape model building. There are several automated methods for solving this problem in 3D, but they usually can only handl...
ICCV
2007
IEEE
14 years 12 months ago
2007
IEEE
We formulate the problem of scene summarization as selecting a set of images that efficiently represents the visual content of a given scene. The ideal summary presents the most i...
ICCV
2007
IEEE
14 years 12 months ago
2007
IEEE
Given an unstructured collection of captioned images of cluttered scenes featuring a variety of objects, our goal is to learn both the names and appearances of the objects. Only a...
ICCV
2007
IEEE
14 years 12 months ago
2007
IEEE
We present a framework for extracting image contours based on geometric and structural consistency among edge element locations and orientations. The paper presents two contributi...
ICCV
2007
IEEE
14 years 12 months ago
2007
IEEE
ICCV
2007
IEEE
14 years 12 months ago
2007
IEEE
Automatically discovering common visual patterns from a collection of images is an interesting but yet challenging task, in part because it is computationally prohibiting. Althoug...
ICCV
2007
IEEE
14 years 12 months ago
2007
IEEE
We present a novel approach to automatically find spatial configurations of local features occurring frequently on instances of a given object class, and rarely on the background....
ICCV
2007
IEEE
14 years 12 months ago
2007
IEEE
In this paper, we present a tracking framework for capturing articulated human motions in real-time, without the need for attaching markers onto the subject's body. This is a...