Better features to track by estimating the tracking convergence region

Zoran Zivkovic, F. van der Heijden

    Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

    7 Citations (Scopus)
    61 Downloads (Pure)

    Abstract

    Reliably tracking key points and textured patches from frame to frame is the basic requirement for many bottom- up computer vision algorithms. The problem of selecting the features that can be tracked well is addressed here. The Lucas- Kcsnade tracking procedure is commonly used. We propose a method so estimate she size of the tracking procedure convergence region for each feature. The features that have a wider convergence region around them should be tracked better by the tracker The size of the convergence region as a new feature goodness measure is compared with the widely accepted She-Tomasi feature selection criteria.
    Original languageEnglish
    Title of host publicationObject recognition supported by user interaction for service robots
    Subtitle of host publicationProceedings 16th International Conference on Pattern Recognition 2002
    EditorsR. Kasturi, D. Laurendeau, C. Suen
    Place of PublicationLos Alamitos, CA
    PublisherIEEE Computer Society
    Pages635-638
    Number of pages4
    Volume4
    ISBN (Print)0-7695-1695-X
    DOIs
    Publication statusPublished - 2002
    Event16th International Conference on Pattern Recognition 2002 - Danvers, Canada, Quebec City, Canada
    Duration: 11 Aug 200215 Aug 2002
    Conference number: 16

    Conference

    Conference16th International Conference on Pattern Recognition 2002
    Abbreviated titleICPR 2002
    CountryCanada
    CityQuebec City
    Period11/08/0215/08/02
    Other11-15 Aug. 2002

    Keywords

    • Bottom-up computer vision algorithms
    • Lucas-Kanade tracking procedure
    • Shi-Tomasi feature selection criteria
    • Feature goodness measure
    • Tracking convergence region
    • Textured patches

    Fingerprint Dive into the research topics of 'Better features to track by estimating the tracking convergence region'. Together they form a unique fingerprint.

    Cite this