Object detection oriented feature pooling for video semantic indexing

Kazuya Ueki, Tetsunori Kobayashi

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

    1 Citation (Scopus)

    Abstract

    We propose a new feature extraction method for video semantic indexing. Conventional methods extract features densely and uniformly across an entire image, whereas the proposed method exploits the object detector to extract features from image windows with high objectness. This feature extraction method focuses on "objects." Therefore, we can eliminate the unnecessary background information, and keep the useful information such as the position, the size, and the aspect ratio of a object. Since these object detection oriented features are complementary to features from entire images, the performance of video semantic indexing can be further improved. Experimental comparisons using large-scale video dataset of the TRECVID benchmark demonstrated that the proposed method substantially improved the performance of video semantic indexing.

    Original languageEnglish
    Title of host publicationVISAPP
    PublisherSciTePress
    Pages44-51
    Number of pages8
    Volume5
    ISBN (Electronic)9789897582264
    Publication statusPublished - 2017 Jan 1
    Event12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2017 - Porto, Portugal
    Duration: 2017 Feb 272017 Mar 1

    Other

    Other12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2017
    CountryPortugal
    CityPorto
    Period17/2/2717/3/1

    Fingerprint

    Semantics
    Feature extraction
    Aspect ratio
    Detectors
    Object detection

    Keywords

    • Convolutional neural network
    • Object detection
    • Video retrieval
    • Video semantic indexing

    ASJC Scopus subject areas

    • Computer Graphics and Computer-Aided Design
    • Computer Vision and Pattern Recognition
    • Artificial Intelligence

    Cite this

    Ueki, K., & Kobayashi, T. (2017). Object detection oriented feature pooling for video semantic indexing. In VISAPP (Vol. 5, pp. 44-51). SciTePress.

    Object detection oriented feature pooling for video semantic indexing. / Ueki, Kazuya; Kobayashi, Tetsunori.

    VISAPP. Vol. 5 SciTePress, 2017. p. 44-51.

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

    Ueki, K & Kobayashi, T 2017, Object detection oriented feature pooling for video semantic indexing. in VISAPP. vol. 5, SciTePress, pp. 44-51, 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2017, Porto, Portugal, 17/2/27.
    Ueki K, Kobayashi T. Object detection oriented feature pooling for video semantic indexing. In VISAPP. Vol. 5. SciTePress. 2017. p. 44-51
    @inproceedings{ee166b0e670e4f02a4bb29540c0eb5c6,
    title = "Object detection oriented feature pooling for video semantic indexing",
    abstract = "We propose a new feature extraction method for video semantic indexing. Conventional methods extract features densely and uniformly across an entire image, whereas the proposed method exploits the object detector to extract features from image windows with high objectness. This feature extraction method focuses on {"}objects.{"} Therefore, we can eliminate the unnecessary background information, and keep the useful information such as the position, the size, and the aspect ratio of a object. Since these object detection oriented features are complementary to features from entire images, the performance of video semantic indexing can be further improved. Experimental comparisons using large-scale video dataset of the TRECVID benchmark demonstrated that the proposed method substantially improved the performance of video semantic indexing.",
    keywords = "Convolutional neural network, Object detection, Video retrieval, Video semantic indexing",
    author = "Kazuya Ueki and Tetsunori Kobayashi",
    year = "2017",
    month = "1",
    day = "1",
    language = "English",
    volume = "5",
    pages = "44--51",
    booktitle = "VISAPP",
    publisher = "SciTePress",

    }

    TY - GEN

    T1 - Object detection oriented feature pooling for video semantic indexing

    AU - Ueki, Kazuya

    AU - Kobayashi, Tetsunori

    PY - 2017/1/1

    Y1 - 2017/1/1

    N2 - We propose a new feature extraction method for video semantic indexing. Conventional methods extract features densely and uniformly across an entire image, whereas the proposed method exploits the object detector to extract features from image windows with high objectness. This feature extraction method focuses on "objects." Therefore, we can eliminate the unnecessary background information, and keep the useful information such as the position, the size, and the aspect ratio of a object. Since these object detection oriented features are complementary to features from entire images, the performance of video semantic indexing can be further improved. Experimental comparisons using large-scale video dataset of the TRECVID benchmark demonstrated that the proposed method substantially improved the performance of video semantic indexing.

    AB - We propose a new feature extraction method for video semantic indexing. Conventional methods extract features densely and uniformly across an entire image, whereas the proposed method exploits the object detector to extract features from image windows with high objectness. This feature extraction method focuses on "objects." Therefore, we can eliminate the unnecessary background information, and keep the useful information such as the position, the size, and the aspect ratio of a object. Since these object detection oriented features are complementary to features from entire images, the performance of video semantic indexing can be further improved. Experimental comparisons using large-scale video dataset of the TRECVID benchmark demonstrated that the proposed method substantially improved the performance of video semantic indexing.

    KW - Convolutional neural network

    KW - Object detection

    KW - Video retrieval

    KW - Video semantic indexing

    UR - http://www.scopus.com/inward/record.url?scp=85047848447&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=85047848447&partnerID=8YFLogxK

    M3 - Conference contribution

    AN - SCOPUS:85047848447

    VL - 5

    SP - 44

    EP - 51

    BT - VISAPP

    PB - SciTePress

    ER -