Adapting Active Shape Models for 3D Segmentation of Tubular Structures in Medical Images

Active Shape Models (ASM) have proven to be an effective approach for image segmentation. In some applications, however, the linear model of gray level appearance around a contour that is used in ASM is not sufficient for accurate boundary localization. Furthermore, the statistical shape model may be too restricted if the training set is limited. This paper describes modifications to both the shape and the appearance model of the original ASM formulation. Shape model flexibility is increased, for tubular objects, by modeling the axis deformation independent of the cross-sectional deformation, and by adding supplementary cylindrical deformation modes. Furthermore, a novel appearance modeling scheme that effectively deals with a highly varying background is developed. In contrast with the conventional ASM approach, the new appearance model is trained on both boundary and non-boundary points, and the probability that a given point belongs to the boundary is estimated non-parametrically. The methods are evaluated on the complex task of segmenting thrombus in abdominal aortic aneurysms (AAA). Shape approximation errors were successfully reduced using the two shape model extensions. Segmentation using the new appearance model significantly outperformed the original ASM scheme; average volume errors are 5.1% and 45% respectively.

[1]  A. Beckett,et al.  AKUFO AND IBARAPA. , 1965, Lancet.

[2]  D. Altman,et al.  STATISTICAL METHODS FOR ASSESSING AGREEMENT BETWEEN TWO METHODS OF CLINICAL MEASUREMENT , 1986, The Lancet.

[3]  Timothy F. Cootes,et al.  Active Shape Models-Their Training and Application , 1995, Comput. Vis. Image Underst..

[4]  Timothy F. Cootes,et al.  Active Appearance Models , 1998, ECCV.

[5]  Milan Sonka,et al.  Segmentation and interpretation of MR brain images. An improved active shape model , 1998, IEEE Transactions on Medical Imaging.

[6]  W. Mali,et al.  Inter- and intraobserver variability of CT measurements obtained after endovascular repair of abdominal aortic aneurysms. , 2000, AJR. American journal of roentgenology.

[7]  Lawrence H. Staib,et al.  Boundary Finding with Prior Shape and Smoothness Models , 2000, IEEE Trans. Pattern Anal. Mach. Intell..

[8]  Anil K. Jain,et al.  Statistical Pattern Recognition: A Review , 2000, IEEE Trans. Pattern Anal. Mach. Intell..

[9]  Milan Sonka,et al.  Object localization and border detection criteria design in edge-based image segmentation: automated learning from examples , 2000, IEEE Transactions on Medical Imaging.

[10]  Milan Sonka,et al.  Active appearance motion models for endocardial contour detection in time sequences of echocardiograms , 2001, SPIE Medical Imaging.

[11]  Sven Loncaric,et al.  3D image analysis of abdominal aortic aneurysm , 2001, Medical Imaging: Image Processing.

[12]  Max A. Viergever,et al.  Active-shape-model-based segmentation of abdominal aortic aneurysms in CTA images , 2002, SPIE Medical Imaging.

[13]  Paul A. Yushkevich,et al.  Multiscale deformable model segmentation and statistical shape analysis using medial descriptions , 2002, IEEE Transactions on Medical Imaging.

[14]  Mads Nielsen,et al.  Computer Vision — ECCV 2002 , 2002, Lecture Notes in Computer Science.

[15]  Alejandro F. Frangi,et al.  Active shape model segmentation with optimal features , 2002, IEEE Transactions on Medical Imaging.

[16]  Jim Graham,et al.  Robust Active Shape Model Search , 2002, ECCV.

[17]  Paul Suetens,et al.  Evaluation of image features and search strategies for segmentation of bone structures in radiographs using Active Shape Models , 2002, Medical Image Anal..