A Likelihood Ratio Classifier for Histogram Features

Raymond Veldhuis, Kiran Raja, Raghavendra Ramachandra

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

    51 Downloads (Pure)

    Abstract

    In a number of classification problems, the features are represented by histograms. Traditionally, histograms are compared by relatively simple distance measures such as the chi-square, the Kullback-Leibler, or the Euclidean distance. This paper proposes a likelihood ratio classifier for histogram features that is optimal in Neyman-Pearson sense. It is based on the assumptions that histograms can be modelled by a multinomial distribution and the bin probabilities of the histograms by a Dirichlet probability den- sity. A simple method to estimate the Dirichlet parameters is included. Feature selection prior to classification improves the classification performance. Classification results are presented on periocular and face data from various datasets. It is shown that the proposed classifier outperforms the chi-square distance measure.
    Original languageEnglish
    Title of host publication2018 IEEE 9th International Conference on Biometrics Theory, Applications and Systems (BTAS)
    Place of PublicationPiscataway, NJ
    PublisherIEEE
    Number of pages8
    ISBN (Electronic)978-1-5386-7180-1, 978-1-5386-7179-5
    ISBN (Print)978-1-5386-7181-8
    DOIs
    Publication statusPublished - 21 Sep 2018
    Event2018 IEEE 9th International Conference on Biometrics Theory, Applications and Systems, BTAS 2018 - Torrance Marriott Redondo Beach, Los Angeles, United States
    Duration: 22 Oct 201825 Oct 2018
    Conference number: 9
    https://www.isi.edu/events/btas2018/home

    Publication series

    NameIEEE International Conference on Biometrics Theory, Applications and Systems (BTAS)
    PublisherIEEE
    Volume2018
    ISSN (Print)2474-9680
    ISSN (Electronic)2474-9699

    Conference

    Conference2018 IEEE 9th International Conference on Biometrics Theory, Applications and Systems, BTAS 2018
    Abbreviated titleBTAS
    CountryUnited States
    CityLos Angeles
    Period22/10/1825/10/18
    Internet address

    Fingerprint Dive into the research topics of 'A Likelihood Ratio Classifier for Histogram Features'. Together they form a unique fingerprint.

  • Cite this

    Veldhuis, R., Raja, K., & Ramachandra, R. (2018). A Likelihood Ratio Classifier for Histogram Features. In 2018 IEEE 9th International Conference on Biometrics Theory, Applications and Systems (BTAS) (IEEE International Conference on Biometrics Theory, Applications and Systems (BTAS); Vol. 2018). Piscataway, NJ: IEEE. https://doi.org/10.1109/BTAS.2018.8698573