• DocumentCode
    3630387
  • Title

    ’Hidden signature’ —- a new solution for on-line verification using DTW

  • Author

    Joanna Putz-Leszczynska;Andrzej Pacut

  • Author_Institution
    NASK, Warsaw, 02 -796, W?wozowa 18, Poland
  • fYear
    2008
  • Firstpage
    162
  • Lastpage
    166
  • Abstract
    The article presents the idea of hidden signature - an artificial signature which can effectively replace the template signature (the best representative of a training set) in signature verification algorithms employing DTW (dynamic time warping). In this paper a few methods for the hidden signature computation are presented with their quality parameters. The main approaches are based on recursive point-by-point averaging and evolutionary algorithms. The hidden signature was tested on the MCYT database, using both genuines and skilled forgeries. The results show that this approach leads to an improvement in verification system parameters.
  • Keywords
    "Handwriting recognition","Evolutionary computation","Testing","Databases","Forgery","Statistics","Heuristic algorithms","Hidden Markov models","Iterative methods","Iterative algorithms"
  • Publisher
    ieee
  • Conference_Titel
    Security Technology, 2008. ICCST 2008. 42nd Annual IEEE International Carnahan Conference on
  • ISSN
    1071-6572
  • Print_ISBN
    978-1-4244-1816-9
  • Electronic_ISBN
    2153-0742
  • Type

    conf

  • DOI
    10.1109/CCST.2008.4751296
  • Filename
    4751296