Extending and applying active appearance models for automated, high precision segmentation in different image modalities

Mikkel Bille Stegmann, Rune Fisker, Bjarne Kjær Ersbøll, Ivar Austvoll (Editor)

    Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

    200 Downloads (Pure)

    Abstract

    In this paper, we present a set of extensions to the deformable template model: Active Appearance Model (AAM) proposed by Cootes et al. AAMs distinguish themselves by learning a priori knowledge through observation of shape and texture variation in a training set. This is used to obtain a compact object class description, which can be employed to rapidly search images for new object instances. The proposed extensions concern enhanced shape representation, handling of homogeneous and heterogeneous textures, refinement optimization using Simulated Annealing and robust statistics. Finally, an initialization scheme is designed thus making the usage of AAMs fully automated. Using these extensions it is demonstrated that AAMs can segment bone structures in radiographs, pork chops in perspective images and the left ventricle in cardiovascular magnetic resonance images in a robust, fast and accurate manner. Subpixel landmark accuracy was obtained in two of the three cases.
    Original languageEnglish
    Title of host publicationProc. 12th Scandinavian Conference on Image Analysis - SCIA 2001, Bergen, Norway
    PublisherNOBIM
    Publication date2001
    Pages90-97
    Publication statusPublished - 2001

    Cite this

    Stegmann, M. B., Fisker, R., Ersbøll, B. K., & Austvoll, I. (Ed.) (2001). Extending and applying active appearance models for automated, high precision segmentation in different image modalities. In Proc. 12th Scandinavian Conference on Image Analysis - SCIA 2001, Bergen, Norway (pp. 90-97). NOBIM.
    Stegmann, Mikkel Bille ; Fisker, Rune ; Ersbøll, Bjarne Kjær ; Austvoll, Ivar (Editor). / Extending and applying active appearance models for automated, high precision segmentation in different image modalities. Proc. 12th Scandinavian Conference on Image Analysis - SCIA 2001, Bergen, Norway. NOBIM, 2001. pp. 90-97
    @inproceedings{b747f01287e44e4089f27c203f7b5928,
    title = "Extending and applying active appearance models for automated, high precision segmentation in different image modalities",
    abstract = "In this paper, we present a set of extensions to the deformable template model: Active Appearance Model (AAM) proposed by Cootes et al. AAMs distinguish themselves by learning a priori knowledge through observation of shape and texture variation in a training set. This is used to obtain a compact object class description, which can be employed to rapidly search images for new object instances. The proposed extensions concern enhanced shape representation, handling of homogeneous and heterogeneous textures, refinement optimization using Simulated Annealing and robust statistics. Finally, an initialization scheme is designed thus making the usage of AAMs fully automated. Using these extensions it is demonstrated that AAMs can segment bone structures in radiographs, pork chops in perspective images and the left ventricle in cardiovascular magnetic resonance images in a robust, fast and accurate manner. Subpixel landmark accuracy was obtained in two of the three cases.",
    author = "Stegmann, {Mikkel Bille} and Rune Fisker and Ersb{\o}ll, {Bjarne Kj{\ae}r} and Ivar Austvoll",
    year = "2001",
    language = "English",
    pages = "90--97",
    booktitle = "Proc. 12th Scandinavian Conference on Image Analysis - SCIA 2001, Bergen, Norway",
    publisher = "NOBIM",

    }

    Stegmann, MB, Fisker, R, Ersbøll, BK & Austvoll, I (ed.) 2001, Extending and applying active appearance models for automated, high precision segmentation in different image modalities. in Proc. 12th Scandinavian Conference on Image Analysis - SCIA 2001, Bergen, Norway. NOBIM, pp. 90-97.

    Extending and applying active appearance models for automated, high precision segmentation in different image modalities. / Stegmann, Mikkel Bille; Fisker, Rune; Ersbøll, Bjarne Kjær; Austvoll, Ivar (Editor).

    Proc. 12th Scandinavian Conference on Image Analysis - SCIA 2001, Bergen, Norway. NOBIM, 2001. p. 90-97.

    Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

    TY - GEN

    T1 - Extending and applying active appearance models for automated, high precision segmentation in different image modalities

    AU - Stegmann, Mikkel Bille

    AU - Fisker, Rune

    AU - Ersbøll, Bjarne Kjær

    A2 - Austvoll, Ivar

    PY - 2001

    Y1 - 2001

    N2 - In this paper, we present a set of extensions to the deformable template model: Active Appearance Model (AAM) proposed by Cootes et al. AAMs distinguish themselves by learning a priori knowledge through observation of shape and texture variation in a training set. This is used to obtain a compact object class description, which can be employed to rapidly search images for new object instances. The proposed extensions concern enhanced shape representation, handling of homogeneous and heterogeneous textures, refinement optimization using Simulated Annealing and robust statistics. Finally, an initialization scheme is designed thus making the usage of AAMs fully automated. Using these extensions it is demonstrated that AAMs can segment bone structures in radiographs, pork chops in perspective images and the left ventricle in cardiovascular magnetic resonance images in a robust, fast and accurate manner. Subpixel landmark accuracy was obtained in two of the three cases.

    AB - In this paper, we present a set of extensions to the deformable template model: Active Appearance Model (AAM) proposed by Cootes et al. AAMs distinguish themselves by learning a priori knowledge through observation of shape and texture variation in a training set. This is used to obtain a compact object class description, which can be employed to rapidly search images for new object instances. The proposed extensions concern enhanced shape representation, handling of homogeneous and heterogeneous textures, refinement optimization using Simulated Annealing and robust statistics. Finally, an initialization scheme is designed thus making the usage of AAMs fully automated. Using these extensions it is demonstrated that AAMs can segment bone structures in radiographs, pork chops in perspective images and the left ventricle in cardiovascular magnetic resonance images in a robust, fast and accurate manner. Subpixel landmark accuracy was obtained in two of the three cases.

    M3 - Article in proceedings

    SP - 90

    EP - 97

    BT - Proc. 12th Scandinavian Conference on Image Analysis - SCIA 2001, Bergen, Norway

    PB - NOBIM

    ER -

    Stegmann MB, Fisker R, Ersbøll BK, Austvoll I, (ed.). Extending and applying active appearance models for automated, high precision segmentation in different image modalities. In Proc. 12th Scandinavian Conference on Image Analysis - SCIA 2001, Bergen, Norway. NOBIM. 2001. p. 90-97