CVPR
2006
IEEE
14 years 6 months ago
2006
IEEE
Many computer vision applications, such as scene analysis and medical image interpretation, are ill-suited for traditional classification where each image can only be associated w...
CVPR
2006
IEEE
14 years 6 months ago
2006
IEEE
For the analysis of images, a deeper understanding of their intrinsic structure is required. This has been obtained for 2D images by means of statistical analysis [15, 18]. Here, ...
CVPR
2006
IEEE
14 years 6 months ago
2006
IEEE
Conventional video summarization methods focus predominantly on summarizing videos along the time axis, such as building a movie trailer. The resulting video trailer tends to reta...
CVPR
2006
IEEE
14 years 6 months ago
2006
IEEE
Recently, the generative modeling approach to video segmentation has been gaining popularity in the computer vision community. For example, the flexible sprites framework has been...
CVPR
2006
IEEE
14 years 6 months ago
2006
IEEE
CVPR
2006
IEEE
14 years 6 months ago
2006
IEEE
It is widely conjectured that the excellent ROC performance of biological vision systems is due in large part to the exploitation of context at each of many levels in a part/whole...
CVPR
2006
IEEE
14 years 6 months ago
2006
IEEE
The solutions to many vision problems involve integrating measurements from multiple sources. Most existing methods rely on a hidden assumption, i.e., these measurements are consi...
CVPR
2006
IEEE
14 years 6 months ago
2006
IEEE
A fast algorithm, Accelerated Kernel Feature Analysis (AKFA), that discovers salient features evidenced in a sample of n unclassified patterns, is presented. Like earlier kernel-b...
CVPR
2006
IEEE
14 years 6 months ago
2006
IEEE
We propose human action detection based on a successive convex matching scheme. Human actions are represented as sequences of postures and specific actions are detected in video b...
CVPR
2006
IEEE
14 years 6 months ago
2006
IEEE
Object detection can be posted as those classification tasks where the rare positive patterns are to be distinguished from the enormous negative patterns. To avoid the danger of m...