Scale-space on image profiles about an object boundary

Sean Ho, Guido Gerig

    Research output: Contribution to journalArticle

    Abstract

    Traditionally, image blurring by diffusion is done in Euclidean space, in an image-based coordinate system. This will blur edges at object boundaries, making segmentation difficult. Geometry-driven diffusion [1] uses a geometric model to steer the blurring, so as to blur along the boundary (to overcome noise) but edge-detect across the object boundary. In this paper, we present a scale-space on image profiles taken about the object boundary, in an object-intrinsic coordinate system. The profiles are sampled from the image in the fashion of Active Shape Models [2], and a scale-space is constructed on the profiles, where diffusion is run only in directions tangent to the boundary. Features from the scale-space are then used to build a statistical model of the image structure about the boundary, trained on a population of images with corresponding geometric models. This statistical image match model can be used in an image segmentation framework. Results are shown in 2D on synthetic and real-world objects; the methods can also be extended to 3D.

    Original languageEnglish (US)
    Pages (from-to)564-575
    Number of pages12
    JournalLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    Volume2695
    StatePublished - 2003

    Fingerprint

    Scale Space
    Statistical Models
    Geometric Model
    Noise
    Image segmentation
    Active Shape Model
    Population
    Geometry
    Profile
    Object
    Image Segmentation
    Tangent line
    Statistical Model
    Euclidean space
    Segmentation

    ASJC Scopus subject areas

    • Computer Science(all)
    • Biochemistry, Genetics and Molecular Biology(all)
    • Theoretical Computer Science

    Cite this

    @article{05b1686ce7354c72b8a558d1391987e4,
    title = "Scale-space on image profiles about an object boundary",
    abstract = "Traditionally, image blurring by diffusion is done in Euclidean space, in an image-based coordinate system. This will blur edges at object boundaries, making segmentation difficult. Geometry-driven diffusion [1] uses a geometric model to steer the blurring, so as to blur along the boundary (to overcome noise) but edge-detect across the object boundary. In this paper, we present a scale-space on image profiles taken about the object boundary, in an object-intrinsic coordinate system. The profiles are sampled from the image in the fashion of Active Shape Models [2], and a scale-space is constructed on the profiles, where diffusion is run only in directions tangent to the boundary. Features from the scale-space are then used to build a statistical model of the image structure about the boundary, trained on a population of images with corresponding geometric models. This statistical image match model can be used in an image segmentation framework. Results are shown in 2D on synthetic and real-world objects; the methods can also be extended to 3D.",
    author = "Sean Ho and Guido Gerig",
    year = "2003",
    language = "English (US)",
    volume = "2695",
    pages = "564--575",
    journal = "Lecture Notes in Computer Science",
    issn = "0302-9743",
    publisher = "Springer Verlag",

    }

    TY - JOUR

    T1 - Scale-space on image profiles about an object boundary

    AU - Ho, Sean

    AU - Gerig, Guido

    PY - 2003

    Y1 - 2003

    N2 - Traditionally, image blurring by diffusion is done in Euclidean space, in an image-based coordinate system. This will blur edges at object boundaries, making segmentation difficult. Geometry-driven diffusion [1] uses a geometric model to steer the blurring, so as to blur along the boundary (to overcome noise) but edge-detect across the object boundary. In this paper, we present a scale-space on image profiles taken about the object boundary, in an object-intrinsic coordinate system. The profiles are sampled from the image in the fashion of Active Shape Models [2], and a scale-space is constructed on the profiles, where diffusion is run only in directions tangent to the boundary. Features from the scale-space are then used to build a statistical model of the image structure about the boundary, trained on a population of images with corresponding geometric models. This statistical image match model can be used in an image segmentation framework. Results are shown in 2D on synthetic and real-world objects; the methods can also be extended to 3D.

    AB - Traditionally, image blurring by diffusion is done in Euclidean space, in an image-based coordinate system. This will blur edges at object boundaries, making segmentation difficult. Geometry-driven diffusion [1] uses a geometric model to steer the blurring, so as to blur along the boundary (to overcome noise) but edge-detect across the object boundary. In this paper, we present a scale-space on image profiles taken about the object boundary, in an object-intrinsic coordinate system. The profiles are sampled from the image in the fashion of Active Shape Models [2], and a scale-space is constructed on the profiles, where diffusion is run only in directions tangent to the boundary. Features from the scale-space are then used to build a statistical model of the image structure about the boundary, trained on a population of images with corresponding geometric models. This statistical image match model can be used in an image segmentation framework. Results are shown in 2D on synthetic and real-world objects; the methods can also be extended to 3D.

    UR - http://www.scopus.com/inward/record.url?scp=35248843225&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=35248843225&partnerID=8YFLogxK

    M3 - Article

    VL - 2695

    SP - 564

    EP - 575

    JO - Lecture Notes in Computer Science

    JF - Lecture Notes in Computer Science

    SN - 0302-9743

    ER -