7928 search results - page 1429 / 1586 » Human-Like Learning Methods for a |
144
click to vote
CVPR
2007
IEEE
16 years 8 months ago
2007
IEEE
This paper presents a framework to automatically detect and recover the occluded facial region. We first derive a Bayesian formulation unifying the occlusion detection and recover...
153
click to vote
CVPR
2007
IEEE
16 years 8 months ago
2007
IEEE
We propose a method for simultaneous detection, localization and segmentation of objects of a known category. We show that this is possible by using segments as features. To this ...
187
click to vote
CVPR
2007
IEEE
16 years 8 months ago
2007
IEEE
Kernel machines have recently been considered as a promising solution for implicit surface modelling. A key challenge of machine learning solutions is how to fit implicit shape mo...
160
click to vote
CVPR
2007
IEEE
16 years 8 months ago
2007
IEEE
Many problems in vision involve the prediction of a class label for each frame in an unsegmented sequence. In this paper, we develop a discriminative framework for simultaneous se...
166
click to vote
CVPR
2007
IEEE
16 years 8 months ago
2007
IEEE
Several authors have noticed that the common representation of images as vectors is sub-optimal. The process of vectorization eliminates spatial relations between some of the near...