Gianfranco Doretto / Publications

Modeling dynamic scenes with active appearance

Doretto, G. and Soatto, S.
Modeling dynamic scenes with active appearance
Technical Report TR040053, UCLA Computer Science Department, 2004.

Download

PDF (265.6kB )  

Abstract

In this work we propose a model for video scenes that contain temporal variability in shape and appearance. We propose a conditionally linear model akin to a dynamic extension of active appearance models. We formulate the problem variationally, and use finite-element methods to compute a numerical solution. We illustrate our model to learn and simulate the shape, appearance, and motion of scenes that exhibit some form of temporal regularity, intended in a statistical sense.

BibTeX

@TechReport{UCLA-CSD-TR040053,
  Title                    = {Modeling dynamic scenes with active appearance},
  Author                   = {Doretto, G. and Soatto, S.},
  Institution              = {UCLA Computer Science Department},
  Year                     = {2004},
  Address                  = {Los Angeles, CA, USA},
  Month                    = dec,
  Number                   = {TR040053},
  Abstract                 = {In this work we propose a model for video scenes that contain temporal variability in shape and appearance. We propose a conditionally linear model akin to a dynamic extension of active appearance models. We formulate the problem variationally, and use finite-element methods to compute a numerical solution. We illustrate our model to learn and simulate the shape, appearance, and motion of scenes that exhibit some form of temporal regularity, intended in a statistical sense.},
  Bib2html_pubtype         = {Tech Reports},
  Bib2html_rescat          = {Dynamic Textures, Shape and Appearance Modeling, Visual Motion Analysis, Image Based Rendering},
  File                     = {UCLA-CSD-TR040053.pdf:doretto\\report\\UCLA-CSD-TR040053.pdf:PDF;UCLA-CSD-TR040053.pdf:doretto\\report\\UCLA-CSD-TR040053.pdf:PDF},
  Owner                    = {doretto}
}