Genetic Programming-Evolved Spatio-Temporal Descriptor for Human Action Recognition

Created by W.Langdon from gp-bibliography.bib Revision:1.4333

  author =       "Li Liu and Ling Shao and Peter Rockett",
  title =        "Genetic Programming-Evolved Spatio-Temporal Descriptor
                 for Human Action Recognition",
  year =         "2012",
  booktitle =    "Proceedings of the British Machine Vision Conference,
                 BMVC 2012",
  editors =      "Richard Bowden and John P. Collomosse and Krystian
  pages =        "18.1--18.12",
  address =      "Surrey, UK",
  month =        sep # " 3-7",
  publisher =    "BMVA Press",
  keywords =     "genetic algorithms, genetic programming",
  ISBN =         "1-901725-46-4",
  bibdate =      "2013-04-24",
  bibsource =    "DBLP,
  URL =          "",
  URL =          "",
  DOI =          "doi:10.5244/C.26.18",
  size =         "12 pages",
  abstract =     "The potential value of human action recognition has
                 led to it becoming one of the most active research
                 subjects in computer vision. In this paper, we propose
                 a novel method to automatically generate low-level
                 spatio-temporal descriptors showing good performance,
                 for high-level human-action recognition tasks. We
                 address this as an optimisation problem using genetic
                 programming (GP), an evolutionary method, which
                 produces the descriptor by combining a set of primitive
                 3D operators. As far as we are aware, this is the first
                 report of using GP for evolving spatio-temporal
                 descriptors for action recognition. In our evolutionary
                 architecture, the average cross-validation
                 classification error calculated using the
                 support-vector machine (SVM) classifier is used as the
                 GP fitness function. We run GP on a mixed dataset
                 combining the KTH and the Weizmann datasets to obtain a
                 promising feature-descriptor solution for action
                 recognition. To demonstrate generalisable, the best
                 descriptor generated so far by GP has also been tested
                 on the IXMAS dataset leading to better accuracies
                 compared with some previous hand-crafted descriptors",
  notes =        "Also known as \cite{conf/bmvc/LiuSR12}",

Genetic Programming entries for Li Liu Ling Shao Peter I Rockett