Novel image and patch manifold models that can exploit the non-linear and non-local redundancies in a dynamic dataset will be introduced. Specifically, the collection of images/patches in the dataset is assumed to be on a smooth manifold. I will introduce novel iterative algorithms to exploit the structure of the data. The use of these algorithms enables implicit motion compensation and motion resolution, and hence is a good alternative to current strategies that perform these operations explicity.
This abstract and the presentation materials are available to members only; a login is required.