Class distance weighted locality preserving projection for automatic age estimation

Kazuya Ueki, Masakazu Miya, Tetsuji Ogawa, Tetsunori Kobayashi

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    6 Citations (Scopus)

    Abstract

    We have developed new dimensionality reduction methods, extended from locality preserving projection (LPP), to estimate age using facial images. LPP seeks a linear transformation matrix such that optimally preserves the neighborhood structure of the data. Our focus has been on expanding LPP by making use of class label information. Specifically, one of our ideas is to assign weights only to the data with close class labels. A local scaling method is used for each class to compute the LPP affinity matrix. Another idea is to assign large weights to two samples with close class labels, i.e., close ages. By doing this, class label information for original data (i.e., age information) can be preserved. We thus call this "class distance weighted linear preserving projection" (CDLPP). Experimental results on a large database showed that CDLPP has more discriminative power than conventional methods such as PCA and LPP.

    Original languageEnglish
    Title of host publicationBTAS 2008 - IEEE 2nd International Conference on Biometrics: Theory, Applications and Systems
    DOIs
    Publication statusPublished - 2008
    EventBTAS 2008 - IEEE 2nd International Conference on Biometrics: Theory, Applications and Systems - Arlington, VA
    Duration: 2008 Sep 292008 Oct 1

    Other

    OtherBTAS 2008 - IEEE 2nd International Conference on Biometrics: Theory, Applications and Systems
    CityArlington, VA
    Period08/9/2908/10/1

    Fingerprint

    Weights and Measures
    Passive Cutaneous Anaphylaxis
    Databases

    ASJC Scopus subject areas

    • Biotechnology

    Cite this

    Ueki, K., Miya, M., Ogawa, T., & Kobayashi, T. (2008). Class distance weighted locality preserving projection for automatic age estimation. In BTAS 2008 - IEEE 2nd International Conference on Biometrics: Theory, Applications and Systems [4699380] https://doi.org/10.1109/BTAS.2008.4699380

    Class distance weighted locality preserving projection for automatic age estimation. / Ueki, Kazuya; Miya, Masakazu; Ogawa, Tetsuji; Kobayashi, Tetsunori.

    BTAS 2008 - IEEE 2nd International Conference on Biometrics: Theory, Applications and Systems. 2008. 4699380.

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Ueki, K, Miya, M, Ogawa, T & Kobayashi, T 2008, Class distance weighted locality preserving projection for automatic age estimation. in BTAS 2008 - IEEE 2nd International Conference on Biometrics: Theory, Applications and Systems., 4699380, BTAS 2008 - IEEE 2nd International Conference on Biometrics: Theory, Applications and Systems, Arlington, VA, 08/9/29. https://doi.org/10.1109/BTAS.2008.4699380
    Ueki K, Miya M, Ogawa T, Kobayashi T. Class distance weighted locality preserving projection for automatic age estimation. In BTAS 2008 - IEEE 2nd International Conference on Biometrics: Theory, Applications and Systems. 2008. 4699380 https://doi.org/10.1109/BTAS.2008.4699380
    Ueki, Kazuya ; Miya, Masakazu ; Ogawa, Tetsuji ; Kobayashi, Tetsunori. / Class distance weighted locality preserving projection for automatic age estimation. BTAS 2008 - IEEE 2nd International Conference on Biometrics: Theory, Applications and Systems. 2008.
    @inproceedings{7af08d6192634be990a3822be7cb5051,
    title = "Class distance weighted locality preserving projection for automatic age estimation",
    abstract = "We have developed new dimensionality reduction methods, extended from locality preserving projection (LPP), to estimate age using facial images. LPP seeks a linear transformation matrix such that optimally preserves the neighborhood structure of the data. Our focus has been on expanding LPP by making use of class label information. Specifically, one of our ideas is to assign weights only to the data with close class labels. A local scaling method is used for each class to compute the LPP affinity matrix. Another idea is to assign large weights to two samples with close class labels, i.e., close ages. By doing this, class label information for original data (i.e., age information) can be preserved. We thus call this {"}class distance weighted linear preserving projection{"} (CDLPP). Experimental results on a large database showed that CDLPP has more discriminative power than conventional methods such as PCA and LPP.",
    author = "Kazuya Ueki and Masakazu Miya and Tetsuji Ogawa and Tetsunori Kobayashi",
    year = "2008",
    doi = "10.1109/BTAS.2008.4699380",
    language = "English",
    isbn = "9781424427307",
    booktitle = "BTAS 2008 - IEEE 2nd International Conference on Biometrics: Theory, Applications and Systems",

    }

    TY - GEN

    T1 - Class distance weighted locality preserving projection for automatic age estimation

    AU - Ueki, Kazuya

    AU - Miya, Masakazu

    AU - Ogawa, Tetsuji

    AU - Kobayashi, Tetsunori

    PY - 2008

    Y1 - 2008

    N2 - We have developed new dimensionality reduction methods, extended from locality preserving projection (LPP), to estimate age using facial images. LPP seeks a linear transformation matrix such that optimally preserves the neighborhood structure of the data. Our focus has been on expanding LPP by making use of class label information. Specifically, one of our ideas is to assign weights only to the data with close class labels. A local scaling method is used for each class to compute the LPP affinity matrix. Another idea is to assign large weights to two samples with close class labels, i.e., close ages. By doing this, class label information for original data (i.e., age information) can be preserved. We thus call this "class distance weighted linear preserving projection" (CDLPP). Experimental results on a large database showed that CDLPP has more discriminative power than conventional methods such as PCA and LPP.

    AB - We have developed new dimensionality reduction methods, extended from locality preserving projection (LPP), to estimate age using facial images. LPP seeks a linear transformation matrix such that optimally preserves the neighborhood structure of the data. Our focus has been on expanding LPP by making use of class label information. Specifically, one of our ideas is to assign weights only to the data with close class labels. A local scaling method is used for each class to compute the LPP affinity matrix. Another idea is to assign large weights to two samples with close class labels, i.e., close ages. By doing this, class label information for original data (i.e., age information) can be preserved. We thus call this "class distance weighted linear preserving projection" (CDLPP). Experimental results on a large database showed that CDLPP has more discriminative power than conventional methods such as PCA and LPP.

    UR - http://www.scopus.com/inward/record.url?scp=67649088888&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=67649088888&partnerID=8YFLogxK

    U2 - 10.1109/BTAS.2008.4699380

    DO - 10.1109/BTAS.2008.4699380

    M3 - Conference contribution

    AN - SCOPUS:67649088888

    SN - 9781424427307

    BT - BTAS 2008 - IEEE 2nd International Conference on Biometrics: Theory, Applications and Systems

    ER -