---

Tensor Methods in Visual Computing

Renato Pajarola, University of Zürich, Svizzera

Aula A, ore 16:00 Palazzo delle Scienze

Abstract:

Tensor decompositions have gained much interest in a wide range of disciplines, ranging from natural sciences and engineering to economics and psychometrics as well as machine learning, as powerful tools to cope with challenges around multidimensional and multivariate data. Tensor decompositions can be viewed as a higher-order generalizations of matrix decompositions and are utilized for various data modeling and analysis tasks. In visual computing, computer graphics, image processing or data visualization, it is common to work with multidimensional data such as, e.g., in the form of image stacks, volume data, vector/tensor fields, videos, or bidirectional texture function (BTF) maps. In this context, tensor decomposition or approximation methods can be explored in particular with respect to compact representation and processing of large-scale and high-dimensional data sets. In this talk we will introduce and focus on the main concepts of tensor decompositions, and give a brief overview of their application to multidimensional visual data.