In this paper, we present a two-layer generative model that incorporates generic middle-level visual knowledge for dense stereo reconstruction. The visual knowledge is represented...
This paper proposes a novel approach to motion capture
from multiple, synchronized video streams, specifically
aimed at recording dense and accurate models of the structure
and ...
Yasutaka Furukawa (University of Washington), Jean...
Blendshapes (linear shape interpolation models) are perhaps the most commonly employed technique in facial animation practice. A major problem in creating blendshape animation is ...
John P. Lewis, Jonathan Mooser, Zhigang Deng, Ulri...
Current two-dimensional image based face recognition systems encounter difficulties with large facial appearance variations due to the pose, illumination and expression changes. ...
— In this paper, we propose an original approach to control camera position and/or lighting conditions in an environment using image gradient information. Our goal is to ensure a...