Class-distance-based discriminant analysis and its application to supervised automatic age estimation

    Research output: Contribution to journalArticle

    Abstract

    We propose a novel method of supervised feature projection called class-distance-based discriminant analysis (CDDA), which is suitable for automatic age estimation (AAE) from facial images. Most methods of supervised feature projection, e.g., Fisher discriminant analysis (FDA) and local Fisher discriminant analysis (LFDA), focus on determining whether two samples belong to the same class (i.e., the same age in AAE) or not. Even if an estimated age is not consistent with the correct age in AAE systems, i.e., the AAE system induces error, smaller errors are better. To treat such characteristics in AAE, CDDA determines between-class separability according to the class distance (i.e., difference in ages); two samples with similar ages are imposed to be close and those with spaced ages are imposed to be far apart. Furthermore, we propose an extension of CDDA called local CDDA (LCDDA), which aims at handling multimodality in samples. Experimental results revealed that CDDA and LCDDA could extract more discriminative features than FDA and LFDA.

    Original languageEnglish
    Pages (from-to)1683-1689
    Number of pages7
    JournalIEICE Transactions on Information and Systems
    VolumeE94-D
    Issue number8
    DOIs
    Publication statusPublished - 2011 Aug

    Fingerprint

    Discriminant analysis

    Keywords

    • Automatic age estimation
    • CDDA
    • Dimensionality reduction
    • FDA
    • LCDDA
    • LFDA

    ASJC Scopus subject areas

    • Electrical and Electronic Engineering
    • Software
    • Artificial Intelligence
    • Hardware and Architecture
    • Computer Vision and Pattern Recognition

    Cite this

    @article{a33d4157159b4882b3d0402f2949e77c,
    title = "Class-distance-based discriminant analysis and its application to supervised automatic age estimation",
    abstract = "We propose a novel method of supervised feature projection called class-distance-based discriminant analysis (CDDA), which is suitable for automatic age estimation (AAE) from facial images. Most methods of supervised feature projection, e.g., Fisher discriminant analysis (FDA) and local Fisher discriminant analysis (LFDA), focus on determining whether two samples belong to the same class (i.e., the same age in AAE) or not. Even if an estimated age is not consistent with the correct age in AAE systems, i.e., the AAE system induces error, smaller errors are better. To treat such characteristics in AAE, CDDA determines between-class separability according to the class distance (i.e., difference in ages); two samples with similar ages are imposed to be close and those with spaced ages are imposed to be far apart. Furthermore, we propose an extension of CDDA called local CDDA (LCDDA), which aims at handling multimodality in samples. Experimental results revealed that CDDA and LCDDA could extract more discriminative features than FDA and LFDA.",
    keywords = "Automatic age estimation, CDDA, Dimensionality reduction, FDA, LCDDA, LFDA",
    author = "Tetsuji Ogawa and Kazuya Ueki and Tetsunori Kobayashi",
    year = "2011",
    month = "8",
    doi = "10.1587/transinf.E94.D.1683",
    language = "English",
    volume = "E94-D",
    pages = "1683--1689",
    journal = "IEICE Transactions on Information and Systems",
    issn = "0916-8532",
    publisher = "Maruzen Co., Ltd/Maruzen Kabushikikaisha",
    number = "8",

    }

    TY - JOUR

    T1 - Class-distance-based discriminant analysis and its application to supervised automatic age estimation

    AU - Ogawa, Tetsuji

    AU - Ueki, Kazuya

    AU - Kobayashi, Tetsunori

    PY - 2011/8

    Y1 - 2011/8

    N2 - We propose a novel method of supervised feature projection called class-distance-based discriminant analysis (CDDA), which is suitable for automatic age estimation (AAE) from facial images. Most methods of supervised feature projection, e.g., Fisher discriminant analysis (FDA) and local Fisher discriminant analysis (LFDA), focus on determining whether two samples belong to the same class (i.e., the same age in AAE) or not. Even if an estimated age is not consistent with the correct age in AAE systems, i.e., the AAE system induces error, smaller errors are better. To treat such characteristics in AAE, CDDA determines between-class separability according to the class distance (i.e., difference in ages); two samples with similar ages are imposed to be close and those with spaced ages are imposed to be far apart. Furthermore, we propose an extension of CDDA called local CDDA (LCDDA), which aims at handling multimodality in samples. Experimental results revealed that CDDA and LCDDA could extract more discriminative features than FDA and LFDA.

    AB - We propose a novel method of supervised feature projection called class-distance-based discriminant analysis (CDDA), which is suitable for automatic age estimation (AAE) from facial images. Most methods of supervised feature projection, e.g., Fisher discriminant analysis (FDA) and local Fisher discriminant analysis (LFDA), focus on determining whether two samples belong to the same class (i.e., the same age in AAE) or not. Even if an estimated age is not consistent with the correct age in AAE systems, i.e., the AAE system induces error, smaller errors are better. To treat such characteristics in AAE, CDDA determines between-class separability according to the class distance (i.e., difference in ages); two samples with similar ages are imposed to be close and those with spaced ages are imposed to be far apart. Furthermore, we propose an extension of CDDA called local CDDA (LCDDA), which aims at handling multimodality in samples. Experimental results revealed that CDDA and LCDDA could extract more discriminative features than FDA and LFDA.

    KW - Automatic age estimation

    KW - CDDA

    KW - Dimensionality reduction

    KW - FDA

    KW - LCDDA

    KW - LFDA

    UR - http://www.scopus.com/inward/record.url?scp=79961039230&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=79961039230&partnerID=8YFLogxK

    U2 - 10.1587/transinf.E94.D.1683

    DO - 10.1587/transinf.E94.D.1683

    M3 - Article

    AN - SCOPUS:79961039230

    VL - E94-D

    SP - 1683

    EP - 1689

    JO - IEICE Transactions on Information and Systems

    JF - IEICE Transactions on Information and Systems

    SN - 0916-8532

    IS - 8

    ER -