Histogram of Oriented Principal Components for Cross-View Action Recognition

Hossein Rahmani, Arif Mahmood, Du Huynh, Ajmal Mian

    Research output: Contribution to journalArticle

    50 Citations (Scopus)

    Abstract

    Existing techniques for 3D action recognition are sensitive to viewpoint variations because they extract features from depth images which are viewpoint dependent. In contrast, we directly process pointclouds for cross-view action recognition from unknown and unseen views. We propose the histogram of oriented principal components (HOPC) descriptor that is robust to noise, viewpoint, scale and action speed variations. At a 3D point, HOPC is computed by projecting the three scaled eigenvectors of the pointcloud within its local spatio-temporal support volume onto the vertices of a regular dodecahedron. HOPC is also used for the detection of spatio-temporal keypoints (STK) in 3D pointcloud sequences so that view-invariant STK descriptors (or Local HOPC descriptors) at these key locations only are used for action recognition. We also propose a global descriptor computed from the normalized spatio-temporal distribution of STKs in 4-D, which we refer to as STK-D. We have evaluated the performance of our proposed descriptors against nine existing techniques on two cross-view and three single-view human action recognition datasets. The experimental results show that our techniques provide significant improvement over state-of-the-art methods.

    Original languageEnglish
    Article number7415989
    Pages (from-to)2430-2443
    Number of pages14
    JournalIEEE Transactions on Pattern Analysis and Machine Intelligence
    Volume38
    Issue number12
    DOIs
    Publication statusPublished - 1 Dec 2016

    Fingerprint

    Action Recognition
    Principal Components
    Eigenvalues and eigenfunctions
    Histogram
    Descriptors
    Point Cloud
    Regular dodecahedron
    Eigenvector
    Unknown
    Invariant
    Dependent
    Experimental Results

    Cite this

    @article{e2e44031d1374f40b8ae8556068f4127,
    title = "Histogram of Oriented Principal Components for Cross-View Action Recognition",
    abstract = "Existing techniques for 3D action recognition are sensitive to viewpoint variations because they extract features from depth images which are viewpoint dependent. In contrast, we directly process pointclouds for cross-view action recognition from unknown and unseen views. We propose the histogram of oriented principal components (HOPC) descriptor that is robust to noise, viewpoint, scale and action speed variations. At a 3D point, HOPC is computed by projecting the three scaled eigenvectors of the pointcloud within its local spatio-temporal support volume onto the vertices of a regular dodecahedron. HOPC is also used for the detection of spatio-temporal keypoints (STK) in 3D pointcloud sequences so that view-invariant STK descriptors (or Local HOPC descriptors) at these key locations only are used for action recognition. We also propose a global descriptor computed from the normalized spatio-temporal distribution of STKs in 4-D, which we refer to as STK-D. We have evaluated the performance of our proposed descriptors against nine existing techniques on two cross-view and three single-view human action recognition datasets. The experimental results show that our techniques provide significant improvement over state-of-the-art methods.",
    keywords = "pointcloud, Spatio-temporal keypoint, view invariance",
    author = "Hossein Rahmani and Arif Mahmood and Du Huynh and Ajmal Mian",
    year = "2016",
    month = "12",
    day = "1",
    doi = "10.1109/TPAMI.2016.2533389",
    language = "English",
    volume = "38",
    pages = "2430--2443",
    journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",
    issn = "0162-8828",
    publisher = "IEEE, Institute of Electrical and Electronics Engineers",
    number = "12",

    }

    Histogram of Oriented Principal Components for Cross-View Action Recognition. / Rahmani, Hossein; Mahmood, Arif; Huynh, Du; Mian, Ajmal.

    In: IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 38, No. 12, 7415989, 01.12.2016, p. 2430-2443.

    Research output: Contribution to journalArticle

    TY - JOUR

    T1 - Histogram of Oriented Principal Components for Cross-View Action Recognition

    AU - Rahmani, Hossein

    AU - Mahmood, Arif

    AU - Huynh, Du

    AU - Mian, Ajmal

    PY - 2016/12/1

    Y1 - 2016/12/1

    N2 - Existing techniques for 3D action recognition are sensitive to viewpoint variations because they extract features from depth images which are viewpoint dependent. In contrast, we directly process pointclouds for cross-view action recognition from unknown and unseen views. We propose the histogram of oriented principal components (HOPC) descriptor that is robust to noise, viewpoint, scale and action speed variations. At a 3D point, HOPC is computed by projecting the three scaled eigenvectors of the pointcloud within its local spatio-temporal support volume onto the vertices of a regular dodecahedron. HOPC is also used for the detection of spatio-temporal keypoints (STK) in 3D pointcloud sequences so that view-invariant STK descriptors (or Local HOPC descriptors) at these key locations only are used for action recognition. We also propose a global descriptor computed from the normalized spatio-temporal distribution of STKs in 4-D, which we refer to as STK-D. We have evaluated the performance of our proposed descriptors against nine existing techniques on two cross-view and three single-view human action recognition datasets. The experimental results show that our techniques provide significant improvement over state-of-the-art methods.

    AB - Existing techniques for 3D action recognition are sensitive to viewpoint variations because they extract features from depth images which are viewpoint dependent. In contrast, we directly process pointclouds for cross-view action recognition from unknown and unseen views. We propose the histogram of oriented principal components (HOPC) descriptor that is robust to noise, viewpoint, scale and action speed variations. At a 3D point, HOPC is computed by projecting the three scaled eigenvectors of the pointcloud within its local spatio-temporal support volume onto the vertices of a regular dodecahedron. HOPC is also used for the detection of spatio-temporal keypoints (STK) in 3D pointcloud sequences so that view-invariant STK descriptors (or Local HOPC descriptors) at these key locations only are used for action recognition. We also propose a global descriptor computed from the normalized spatio-temporal distribution of STKs in 4-D, which we refer to as STK-D. We have evaluated the performance of our proposed descriptors against nine existing techniques on two cross-view and three single-view human action recognition datasets. The experimental results show that our techniques provide significant improvement over state-of-the-art methods.

    KW - pointcloud

    KW - Spatio-temporal keypoint

    KW - view invariance

    UR - http://www.scopus.com/inward/record.url?scp=84995584051&partnerID=8YFLogxK

    U2 - 10.1109/TPAMI.2016.2533389

    DO - 10.1109/TPAMI.2016.2533389

    M3 - Article

    VL - 38

    SP - 2430

    EP - 2443

    JO - IEEE Transactions on Pattern Analysis and Machine Intelligence

    JF - IEEE Transactions on Pattern Analysis and Machine Intelligence

    SN - 0162-8828

    IS - 12

    M1 - 7415989

    ER -