Tensor decomposition and application in image classification with histogram of oriented gradients

    Research output: Contribution to journalArticle

    10 Citations (Scopus)

    Abstract

    In the field of visual data mining, Histogram of Oriented Gradients (HOG) and its variants have been widely used. The speed and ability to extract image features that are robust against many types of distortions such as scaling, orientation, affine and illumination that HOG offers have made it a popular choice for the task of detecting images in scenes for classification. However, the high dimensionality nature of HOG descriptors (features), usually in the order of multiple thousands of them per image, would require careful consideration in place to achieve accurate and timely categorization of objects within images. This work explores the possibility of processing HOG features as tensors, or multi-dimensional arrays. A direct result of that is tensor decomposition techniques such as canonical polyadic (CP) decomposition performed on the high-order HOG tensors as the mean for dimensionality reduction by filtering. This work focuses on the impact of this approach on both accuracy and efficiency, comparing it against the standard practice of processing HOG features. Validating with the Caltech-101 dataset, the results achieved with artificial neural network (ANN) classification indicate that the proposed method not only improves the overall system performance, it also achieves the edge in accuracy by a notable margin.
    Original languageEnglish
    Pages (from-to)38-45
    Number of pages8
    JournalNeurocomputing
    Volume165
    Issue number5
    DOIs
    Publication statusPublished - 2015

    Fingerprint

    Image classification
    Tensors
    Decomposition
    Aptitude
    Data Mining
    Visual Fields
    Lighting
    Processing
    Efficiency
    Data mining
    Neural networks
    Datasets

    Cite this

    @article{035ad8979196454eab2ebd77ca967eaa,
    title = "Tensor decomposition and application in image classification with histogram of oriented gradients",
    abstract = "In the field of visual data mining, Histogram of Oriented Gradients (HOG) and its variants have been widely used. The speed and ability to extract image features that are robust against many types of distortions such as scaling, orientation, affine and illumination that HOG offers have made it a popular choice for the task of detecting images in scenes for classification. However, the high dimensionality nature of HOG descriptors (features), usually in the order of multiple thousands of them per image, would require careful consideration in place to achieve accurate and timely categorization of objects within images. This work explores the possibility of processing HOG features as tensors, or multi-dimensional arrays. A direct result of that is tensor decomposition techniques such as canonical polyadic (CP) decomposition performed on the high-order HOG tensors as the mean for dimensionality reduction by filtering. This work focuses on the impact of this approach on both accuracy and efficiency, comparing it against the standard practice of processing HOG features. Validating with the Caltech-101 dataset, the results achieved with artificial neural network (ANN) classification indicate that the proposed method not only improves the overall system performance, it also achieves the edge in accuracy by a notable margin.",
    keywords = "image classification, histogram of oriented gradients, tensor decomposition",
    author = "Dat TRAN and Wanli MA and Tan VO",
    year = "2015",
    doi = "10.1016/j.neucom.2014.06.093",
    language = "English",
    volume = "165",
    pages = "38--45",
    journal = "Neurocomputing: an international journal",
    issn = "0925-2312",
    publisher = "Elsevier",
    number = "5",

    }

    Tensor decomposition and application in image classification with histogram of oriented gradients. / TRAN, Dat; MA, Wanli; VO, Tan.

    In: Neurocomputing, Vol. 165, No. 5, 2015, p. 38-45.

    Research output: Contribution to journalArticle

    TY - JOUR

    T1 - Tensor decomposition and application in image classification with histogram of oriented gradients

    AU - TRAN, Dat

    AU - MA, Wanli

    AU - VO, Tan

    PY - 2015

    Y1 - 2015

    N2 - In the field of visual data mining, Histogram of Oriented Gradients (HOG) and its variants have been widely used. The speed and ability to extract image features that are robust against many types of distortions such as scaling, orientation, affine and illumination that HOG offers have made it a popular choice for the task of detecting images in scenes for classification. However, the high dimensionality nature of HOG descriptors (features), usually in the order of multiple thousands of them per image, would require careful consideration in place to achieve accurate and timely categorization of objects within images. This work explores the possibility of processing HOG features as tensors, or multi-dimensional arrays. A direct result of that is tensor decomposition techniques such as canonical polyadic (CP) decomposition performed on the high-order HOG tensors as the mean for dimensionality reduction by filtering. This work focuses on the impact of this approach on both accuracy and efficiency, comparing it against the standard practice of processing HOG features. Validating with the Caltech-101 dataset, the results achieved with artificial neural network (ANN) classification indicate that the proposed method not only improves the overall system performance, it also achieves the edge in accuracy by a notable margin.

    AB - In the field of visual data mining, Histogram of Oriented Gradients (HOG) and its variants have been widely used. The speed and ability to extract image features that are robust against many types of distortions such as scaling, orientation, affine and illumination that HOG offers have made it a popular choice for the task of detecting images in scenes for classification. However, the high dimensionality nature of HOG descriptors (features), usually in the order of multiple thousands of them per image, would require careful consideration in place to achieve accurate and timely categorization of objects within images. This work explores the possibility of processing HOG features as tensors, or multi-dimensional arrays. A direct result of that is tensor decomposition techniques such as canonical polyadic (CP) decomposition performed on the high-order HOG tensors as the mean for dimensionality reduction by filtering. This work focuses on the impact of this approach on both accuracy and efficiency, comparing it against the standard practice of processing HOG features. Validating with the Caltech-101 dataset, the results achieved with artificial neural network (ANN) classification indicate that the proposed method not only improves the overall system performance, it also achieves the edge in accuracy by a notable margin.

    KW - image classification

    KW - histogram of oriented gradients

    KW - tensor decomposition

    U2 - 10.1016/j.neucom.2014.06.093

    DO - 10.1016/j.neucom.2014.06.093

    M3 - Article

    VL - 165

    SP - 38

    EP - 45

    JO - Neurocomputing: an international journal

    JF - Neurocomputing: an international journal

    SN - 0925-2312

    IS - 5

    ER -