A fast algorithm for creating a compact and discriminative visual codebook

Lei Wang*, Luping Zhou, Chunhua Shen

*Corresponding author for this work

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

    16 Citations (Scopus)

    Abstract

    In patch-based object recognition, using a compact visual codebook can boost computational efficiency and reduce memory cost. Nevertheless, compared with a large-sized codebook, it also risks the loss of discriminative power. Moreover, creating a compact visual codebook can be very time-consuming, especially when the number of initial visual words is large. In this paper, to minimize its loss of discriminative power, we propose an approach to build a compact visual codebook by maximally preserving the separability of the object classes. Furthermore, a fast algorithm is designed to accomplish this task effortlessly, which can hierarchically merge 10,000 visual words down to 2 in ninety seconds. Experimental study shows that the compact visual codebook created in this way can achieve excellent classification performance even after a considerable reduction in size.

    Original languageEnglish
    Title of host publicationComputer Vision - ECCV 2008 - 10th European Conference on Computer Vision, Proceedings
    PublisherSpringer Verlag
    Pages719-732
    Number of pages14
    EditionPART 4
    ISBN (Print)3540886923, 9783540886921
    DOIs
    Publication statusPublished - 2008
    Event10th European Conference on Computer Vision, ECCV 2008 - Marseille, France
    Duration: 12 Oct 200818 Oct 2008

    Publication series

    NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    NumberPART 4
    Volume5305 LNCS
    ISSN (Print)0302-9743
    ISSN (Electronic)1611-3349

    Conference

    Conference10th European Conference on Computer Vision, ECCV 2008
    Country/TerritoryFrance
    CityMarseille
    Period12/10/0818/10/08

    Fingerprint

    Dive into the research topics of 'A fast algorithm for creating a compact and discriminative visual codebook'. Together they form a unique fingerprint.

    Cite this