Our objective is to learn representations for the shape and the appearance of moving (dynamic) objects that support tasks such as synthesis, pose recovery, reconstruction, and tracking. In this paper, we introduce a framework that aim to learn a landmark-free correspondence-free global representations of dynamic appearance manifolds. We use nonlinear dimensionality reduction to achieve an embed...

Having a good description of an object’s appearance is crucial for good object tracking. However, modeling the whole appearance of an object is difficult because of the high dimensional and nonlinear character of the appearance. To tackle the first problem we apply nonlinear dimensionality reduction approaches on multiple views of an object in order to extract the appearance manifold of the obj...

Stochastic analysis of random heterogeneous media provides information of significance only if realistic input models of the topology and material property variations are used. This work introduces a framework to construct such input stochastic models for the topology, thermal diffusivity and permeability variations in heterogeneous media using a data-driven strategy. Given a set of microstruct...

Traditional manifold learning algorithms often bear an assumption that the local neighborhood of any point on embedded manifold is roughly equal to the tangent space at that point without considering the curvature. The curvature indifferent way of manifold processing often makes traditional dimension reduction poorly neighborhood preserving. To overcome this drawback we propose a new algorithm ...

The locally linear embedding (LLE) algorithm has recently emerged as a promising technique for nonlinear dimensionality reduction of high-dimensional data. One of its advantages over many similar methods is that only one parameter has to be defined, but no guidance was yet given how to choose it. We propose a hierarchical method for automatic selection of an optimal parameter value. Our approac...