Possibilities of zero-error recognition by dissimilarity representations

RPW Duin, EM Pekalska

    Research output: Chapter in Book/Conference proceedings/Edited volumeConference contributionScientificpeer-review

    Abstract

    Feature based approaches to pattern recognition suffer from the fact that feature representations of different classes of objects may overlap. This is the consequence of reducing the description of an object to a feature vector. As a result an error free recognition system is even asymptotically (for infinite training sizes) impossible. In this paper it is argued that this limitation does not hold for dissimilarity based representations. Suggestions are made how this may be exploited in practice.
    Original languageUndefined/Unknown
    Title of host publicationPattern Recognition in Information Systems, Proceedings
    EditorsJ.M. Iñesta, L Micó
    Place of PublicationSetubal, Portugal
    PublisherICEIS Press
    Pages20-32
    Number of pages13
    ISBN (Print)972-98050-8-3
    Publication statusPublished - 2002
    Event2nd International Workshop on PRIS, Alicante, Spain - Setubal, Portugal
    Duration: 1 Apr 20021 Apr 2002

    Publication series

    Name
    PublisherICEIS Press

    Conference

    Conference2nd International Workshop on PRIS, Alicante, Spain
    Period1/04/021/04/02

    Keywords

    • conference contrib. refereed
    • Conf.proc. > 3 pag

    Cite this