Fast Stochastic Optimization for Articulated Structure Tracking

M. Bray, E. Koller-Meier, N. N. Schraudolph, and L. Van Gool. Fast Stochastic Optimization for Articulated Structure Tracking. Image and Vision Computing, 25(3):352–364, 2007.
Earlier version

Download

1.3MB	823.7kB	1.8MB

Abstract

Recently, an optimization approach for fast visual tracking of articulated structures based on stochastic meta-descent (SMD) has been presented. SMD is a gradient descent with local step size adaptation that combines rapid convergence with excellent scalability. Stochastic sampling helps to avoid local minima in the optimization process. We have extended the SMD algorithm with new features for fast and accurate tracking by adapting the different step sizes between as well as within video frames and by introducing a robust cost function, which incorporates both depths and surface orientations. The advantages of the resulting tracker over state-of-the-art methods are supported through 3D hand tracking experiments. A realistic deformable hand model reinforces the accuracy of our tracker.

BibTeX Entry

@article{BraKolSchVan07,
     author = {Matthieu Bray and Esther Koller-Meier and
               Nicol N. Schraudolph and Luc Van~Gool},
      title = {\href{http://nic.schraudolph.org/pubs/BraKolSchVan07.pdf}{Fast
               Stochastic Optimization for Articulated Structure Tracking}},
      pages = {352--364},
    journal = {Image and Vision Computing},
     volume =  25,
     number =  3,
       year =  2007,
   b2h_type = {Journal Papers},
  b2h_topic = {>Stochastic Meta-Descent, Computer Vision},
   b2h_note = {<a href="b2hd-BraKolSchVan04.html">Earlier version</a>},
   abstract = {
    Recently, an optimization approach for fast visual tracking of
    articulated structures based on stochastic meta-descent (SMD)
    has been presented. SMD is a gradient descent with local step
    size adaptation that combines rapid convergence with excellent
    scalability. Stochastic sampling helps to avoid local minima
    in the optimization process. We have extended the SMD algorithm
    with new features for fast and accurate tracking by adapting
    the different step sizes between as well as within video frames
    and by introducing a robust cost function, which incorporates
    both depths and surface orientations. The advantages of the
    resulting tracker over state-of-the-art methods are supported
    through 3D hand tracking experiments. A realistic deformable
    hand model reinforces the accuracy of our tracker.
}}