This paper presents a novel method for surface recovery from discrete 3D point data sets. In order to produce improved reconstruction results, the algorithm presented in this paper...
Algorithms for nonlinear dimensionality reduction (NLDR) find meaningful hidden low-dimensional structures in a high-dimensional space. Current algorithms for NLDR are Isomaps, Loc...
We propose a generative model that codes the geometry and appearance of generic visual object categories as a loose hierarchy of parts, with probabilistic spatial relations linkin...
Locally Orderless Tracking (LOT) is a visual tracking algorithm that automatically estimates the amount of local (dis)order in the object. This lets the tracker specialize in both...
Shaul Oron, Aharon Bar-Hillel, Dan Levi, Shai Avid...
Color processing imposes a new constraint on stereo vision algorithms: The assumption of constant color on object surfaces used to align local correlation windows with object bound...