Level-set evolution with region competition: Automatic 3-D segmentation of brain tumors

Sean Ho, Elizabeth Bullitt, Guido Gerig

    Research output: Chapter in Book/Report/Conference proceedingChapter

    Abstract

    We develop a new method for automatic segmentation of anatomical structures from volumetric medical images. Driving application is tumor segmentation from 3-D MRIs, which is known to be a very challenging problem due to the variability of tumor geometry and intensity patterns. Level-set snakes offer significant advantages over conventional statistical classification and mathematical morphology, however snakes with constant propagation need careful initialization and can leak through weak or missing boundary parts. Our region competition method overcomes these problems by modulating the propagation term with a signed local statistical force, leading to a stable solution. A pre- vs. post-contrast difference image is used to calculate probabilities for background and tumor regions, with a mixture-modelling fit of the histogram. Preliminary results on five cases with significant shape and intensity variability demonstrate that the new method might become a powerful and efficient tool for the clinic. Validity is demonstrated by comparison with manual expert segmentation.

    Original languageEnglish (US)
    Title of host publicationProceedings - International Conference on Pattern Recognition
    Pages532-535
    Number of pages4
    Volume16
    Edition1
    StatePublished - 2002

    Fingerprint

    Tumors
    Brain
    Mathematical morphology
    Magnetic resonance imaging
    Geometry

    ASJC Scopus subject areas

    • Electrical and Electronic Engineering
    • Computer Vision and Pattern Recognition
    • Hardware and Architecture

    Cite this

    Ho, S., Bullitt, E., & Gerig, G. (2002). Level-set evolution with region competition: Automatic 3-D segmentation of brain tumors. In Proceedings - International Conference on Pattern Recognition (1 ed., Vol. 16, pp. 532-535)

    Level-set evolution with region competition : Automatic 3-D segmentation of brain tumors. / Ho, Sean; Bullitt, Elizabeth; Gerig, Guido.

    Proceedings - International Conference on Pattern Recognition. Vol. 16 1. ed. 2002. p. 532-535.

    Research output: Chapter in Book/Report/Conference proceedingChapter

    Ho, S, Bullitt, E & Gerig, G 2002, Level-set evolution with region competition: Automatic 3-D segmentation of brain tumors. in Proceedings - International Conference on Pattern Recognition. 1 edn, vol. 16, pp. 532-535.
    Ho S, Bullitt E, Gerig G. Level-set evolution with region competition: Automatic 3-D segmentation of brain tumors. In Proceedings - International Conference on Pattern Recognition. 1 ed. Vol. 16. 2002. p. 532-535
    Ho, Sean ; Bullitt, Elizabeth ; Gerig, Guido. / Level-set evolution with region competition : Automatic 3-D segmentation of brain tumors. Proceedings - International Conference on Pattern Recognition. Vol. 16 1. ed. 2002. pp. 532-535
    @inbook{d3c8a5357866489cbb955a17a3ead328,
    title = "Level-set evolution with region competition: Automatic 3-D segmentation of brain tumors",
    abstract = "We develop a new method for automatic segmentation of anatomical structures from volumetric medical images. Driving application is tumor segmentation from 3-D MRIs, which is known to be a very challenging problem due to the variability of tumor geometry and intensity patterns. Level-set snakes offer significant advantages over conventional statistical classification and mathematical morphology, however snakes with constant propagation need careful initialization and can leak through weak or missing boundary parts. Our region competition method overcomes these problems by modulating the propagation term with a signed local statistical force, leading to a stable solution. A pre- vs. post-contrast difference image is used to calculate probabilities for background and tumor regions, with a mixture-modelling fit of the histogram. Preliminary results on five cases with significant shape and intensity variability demonstrate that the new method might become a powerful and efficient tool for the clinic. Validity is demonstrated by comparison with manual expert segmentation.",
    author = "Sean Ho and Elizabeth Bullitt and Guido Gerig",
    year = "2002",
    language = "English (US)",
    volume = "16",
    pages = "532--535",
    booktitle = "Proceedings - International Conference on Pattern Recognition",
    edition = "1",

    }

    TY - CHAP

    T1 - Level-set evolution with region competition

    T2 - Automatic 3-D segmentation of brain tumors

    AU - Ho, Sean

    AU - Bullitt, Elizabeth

    AU - Gerig, Guido

    PY - 2002

    Y1 - 2002

    N2 - We develop a new method for automatic segmentation of anatomical structures from volumetric medical images. Driving application is tumor segmentation from 3-D MRIs, which is known to be a very challenging problem due to the variability of tumor geometry and intensity patterns. Level-set snakes offer significant advantages over conventional statistical classification and mathematical morphology, however snakes with constant propagation need careful initialization and can leak through weak or missing boundary parts. Our region competition method overcomes these problems by modulating the propagation term with a signed local statistical force, leading to a stable solution. A pre- vs. post-contrast difference image is used to calculate probabilities for background and tumor regions, with a mixture-modelling fit of the histogram. Preliminary results on five cases with significant shape and intensity variability demonstrate that the new method might become a powerful and efficient tool for the clinic. Validity is demonstrated by comparison with manual expert segmentation.

    AB - We develop a new method for automatic segmentation of anatomical structures from volumetric medical images. Driving application is tumor segmentation from 3-D MRIs, which is known to be a very challenging problem due to the variability of tumor geometry and intensity patterns. Level-set snakes offer significant advantages over conventional statistical classification and mathematical morphology, however snakes with constant propagation need careful initialization and can leak through weak or missing boundary parts. Our region competition method overcomes these problems by modulating the propagation term with a signed local statistical force, leading to a stable solution. A pre- vs. post-contrast difference image is used to calculate probabilities for background and tumor regions, with a mixture-modelling fit of the histogram. Preliminary results on five cases with significant shape and intensity variability demonstrate that the new method might become a powerful and efficient tool for the clinic. Validity is demonstrated by comparison with manual expert segmentation.

    UR - http://www.scopus.com/inward/record.url?scp=33751563521&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=33751563521&partnerID=8YFLogxK

    M3 - Chapter

    AN - SCOPUS:33751563521

    VL - 16

    SP - 532

    EP - 535

    BT - Proceedings - International Conference on Pattern Recognition

    ER -