Detection of Cast Shadows in Surveillance Applications

Søren G. Erbou, Helge Bjarne Dissing Sørensen, Bjarne Stage

    Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearch

    292 Downloads (Pure)

    Abstract

    Cast shadows from moving objects reduce the general ability of robust classification and tracking of these objects, in outdoor surveillance applications. A method for segmentation of cast shadows is proposed, combining statistical features with a new similarity feature, derived from a physics-based model. The new method is compared to a reference method, and found to improve performance significantly, based on a test set of real-world examples.
    Original languageEnglish
    Title of host publicationDen 14. Danske Konference i Mønstergenkendelse og Billedanalyse
    PublisherInformatics and Mathematical Modelling, Technical University of Denmark, DTU
    Publication date2005
    Publication statusPublished - 2005

    Cite this

    Erbou, S. G., Sørensen, H. B. D., & Stage, B. (2005). Detection of Cast Shadows in Surveillance Applications. In Den 14. Danske Konference i Mønstergenkendelse og Billedanalyse Informatics and Mathematical Modelling, Technical University of Denmark, DTU.
    Erbou, Søren G. ; Sørensen, Helge Bjarne Dissing ; Stage, Bjarne. / Detection of Cast Shadows in Surveillance Applications. Den 14. Danske Konference i Mønstergenkendelse og Billedanalyse. Informatics and Mathematical Modelling, Technical University of Denmark, DTU, 2005.
    @inproceedings{80687bf93e604b1c8a53e0265b76a128,
    title = "Detection of Cast Shadows in Surveillance Applications",
    abstract = "Cast shadows from moving objects reduce the general ability of robust classification and tracking of these objects, in outdoor surveillance applications. A method for segmentation of cast shadows is proposed, combining statistical features with a new similarity feature, derived from a physics-based model. The new method is compared to a reference method, and found to improve performance significantly, based on a test set of real-world examples.",
    keywords = "Video Surveillance, Shadow Removal, Cast Shadows",
    author = "Erbou, {S{\o}ren G.} and S{\o}rensen, {Helge Bjarne Dissing} and Bjarne Stage",
    year = "2005",
    language = "English",
    booktitle = "Den 14. Danske Konference i M{\o}nstergenkendelse og Billedanalyse",
    publisher = "Informatics and Mathematical Modelling, Technical University of Denmark, DTU",

    }

    Erbou, SG, Sørensen, HBD & Stage, B 2005, Detection of Cast Shadows in Surveillance Applications. in Den 14. Danske Konference i Mønstergenkendelse og Billedanalyse. Informatics and Mathematical Modelling, Technical University of Denmark, DTU.

    Detection of Cast Shadows in Surveillance Applications. / Erbou, Søren G.; Sørensen, Helge Bjarne Dissing; Stage, Bjarne.

    Den 14. Danske Konference i Mønstergenkendelse og Billedanalyse. Informatics and Mathematical Modelling, Technical University of Denmark, DTU, 2005.

    Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearch

    TY - GEN

    T1 - Detection of Cast Shadows in Surveillance Applications

    AU - Erbou, Søren G.

    AU - Sørensen, Helge Bjarne Dissing

    AU - Stage, Bjarne

    PY - 2005

    Y1 - 2005

    N2 - Cast shadows from moving objects reduce the general ability of robust classification and tracking of these objects, in outdoor surveillance applications. A method for segmentation of cast shadows is proposed, combining statistical features with a new similarity feature, derived from a physics-based model. The new method is compared to a reference method, and found to improve performance significantly, based on a test set of real-world examples.

    AB - Cast shadows from moving objects reduce the general ability of robust classification and tracking of these objects, in outdoor surveillance applications. A method for segmentation of cast shadows is proposed, combining statistical features with a new similarity feature, derived from a physics-based model. The new method is compared to a reference method, and found to improve performance significantly, based on a test set of real-world examples.

    KW - Video Surveillance

    KW - Shadow Removal

    KW - Cast Shadows

    M3 - Article in proceedings

    BT - Den 14. Danske Konference i Mønstergenkendelse og Billedanalyse

    PB - Informatics and Mathematical Modelling, Technical University of Denmark, DTU

    ER -

    Erbou SG, Sørensen HBD, Stage B. Detection of Cast Shadows in Surveillance Applications. In Den 14. Danske Konference i Mønstergenkendelse og Billedanalyse. Informatics and Mathematical Modelling, Technical University of Denmark, DTU. 2005