k-fold Subsampling based Sequential Backward Feature Elimination

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

    Published

    Standard

    k-fold Subsampling based Sequential Backward Feature Elimination. / Park, Jeonghwan; Li, Kang; Zhou, Huiyu.

    Proceedings of the 5th International Conference on Pattern Recognition Applications and Methods. SciTePress, 2016. p. 423-430.

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

    Harvard

    Park, J, Li, K & Zhou, H 2016, k-fold Subsampling based Sequential Backward Feature Elimination. in Proceedings of the 5th International Conference on Pattern Recognition Applications and Methods. SciTePress, pp. 423-430, International Conference on Pattern Recognition Applications and Methods (ICPRAM), Rome, Italy, 24/02/2016. https://doi.org/10.5220/0005688804230430

    APA

    Park, J., Li, K., & Zhou, H. (2016). k-fold Subsampling based Sequential Backward Feature Elimination. In Proceedings of the 5th International Conference on Pattern Recognition Applications and Methods (pp. 423-430). SciTePress. https://doi.org/10.5220/0005688804230430

    Vancouver

    Park J, Li K, Zhou H. k-fold Subsampling based Sequential Backward Feature Elimination. In Proceedings of the 5th International Conference on Pattern Recognition Applications and Methods. SciTePress. 2016. p. 423-430 https://doi.org/10.5220/0005688804230430

    Author

    Park, Jeonghwan ; Li, Kang ; Zhou, Huiyu. / k-fold Subsampling based Sequential Backward Feature Elimination. Proceedings of the 5th International Conference on Pattern Recognition Applications and Methods. SciTePress, 2016. pp. 423-430

    Bibtex

    @inproceedings{d06be7ae5b9542309b16137aaaea6372,
    title = "k-fold Subsampling based Sequential Backward Feature Elimination",
    abstract = "We present a new wrapper feature selection algorithm for human detection. This algorithm is a hybrid featureselection approach combining the benefits of filter and wrapper methods. It allows the selection of an optimalfeature vector that well represents the shapes of the subjects in the images. In detail, the proposed featureselection algorithm adopts the k-fold subsampling and sequential backward elimination approach, while thestandard linear support vector machine (SVM) is used as the classifier for human detection. We apply theproposed algorithm to the publicly accessible INRIA and ETH pedestrian full image datasets with the PASCALVOC evaluation criteria. Compared to other state of the arts algorithms, our feature selection based approachcan improve the detection speed of the SVM classifier by over 50{\%} with up to 2{\%} better detection accuracy.Our algorithm also outperforms the equivalent systems introduced in the deformable part model approach witharound 9{\%} improvement in the detection accuracy",
    author = "Jeonghwan Park and Kang Li and Huiyu Zhou",
    year = "2016",
    month = "2",
    doi = "10.5220/0005688804230430",
    language = "English",
    pages = "423--430",
    booktitle = "Proceedings of the 5th International Conference on Pattern Recognition Applications and Methods",
    publisher = "SciTePress",

    }

    RIS

    TY - GEN

    T1 - k-fold Subsampling based Sequential Backward Feature Elimination

    AU - Park, Jeonghwan

    AU - Li, Kang

    AU - Zhou, Huiyu

    PY - 2016/2

    Y1 - 2016/2

    N2 - We present a new wrapper feature selection algorithm for human detection. This algorithm is a hybrid featureselection approach combining the benefits of filter and wrapper methods. It allows the selection of an optimalfeature vector that well represents the shapes of the subjects in the images. In detail, the proposed featureselection algorithm adopts the k-fold subsampling and sequential backward elimination approach, while thestandard linear support vector machine (SVM) is used as the classifier for human detection. We apply theproposed algorithm to the publicly accessible INRIA and ETH pedestrian full image datasets with the PASCALVOC evaluation criteria. Compared to other state of the arts algorithms, our feature selection based approachcan improve the detection speed of the SVM classifier by over 50% with up to 2% better detection accuracy.Our algorithm also outperforms the equivalent systems introduced in the deformable part model approach witharound 9% improvement in the detection accuracy

    AB - We present a new wrapper feature selection algorithm for human detection. This algorithm is a hybrid featureselection approach combining the benefits of filter and wrapper methods. It allows the selection of an optimalfeature vector that well represents the shapes of the subjects in the images. In detail, the proposed featureselection algorithm adopts the k-fold subsampling and sequential backward elimination approach, while thestandard linear support vector machine (SVM) is used as the classifier for human detection. We apply theproposed algorithm to the publicly accessible INRIA and ETH pedestrian full image datasets with the PASCALVOC evaluation criteria. Compared to other state of the arts algorithms, our feature selection based approachcan improve the detection speed of the SVM classifier by over 50% with up to 2% better detection accuracy.Our algorithm also outperforms the equivalent systems introduced in the deformable part model approach witharound 9% improvement in the detection accuracy

    UR - http://www.icpram.org/?y=2016

    U2 - 10.5220/0005688804230430

    DO - 10.5220/0005688804230430

    M3 - Conference contribution

    SP - 423

    EP - 430

    BT - Proceedings of the 5th International Conference on Pattern Recognition Applications and Methods

    PB - SciTePress

    ER -

    Download

    Download as: RIS