Dynamical Systems in Video Analysis
Gianfranco Doretto*,1, Avinash Ravichandran†,1, René Vidal‡ and Stefano Soatto†, *West Virginia University, USA, †University of California, Los Angeles, USA, ‡Johns Hopkins University, USA, email@example.com, firstname.lastname@example.org, email@example.com, firstname.lastname@example.org
This chapter describes how the class of linear dynamical system (LDS) models can be used for representing and analyzing video signals. Given a sequence of images, it is shown under what conditions it can be modeled with an LDS, and how it is possible to estimate the model parameters (system identification). Video data become LDSs, and establishing a way to compare them enables several higher-level vision tasks. ...