A Computer-Aided Design System for Volume Image Segmentation

Sponsored by: Kettering Medical Center and Wright Laboratories

PI: Ardy Goshtasby, Wright State University

Co-PI: Martin Satter, Wallace-Kettering Neuroscience Institute

Students: Marcel Jackowski, Shawn Cheng, and Anurag Jain


Abstract.

Interactive tools for segmenting 2-D and 3-D images are being developed. These tools will allow a user to quickly revise a segmentation result obtained from an automatic method. A thresholding technique is described that finds a unique threshold value for each homogeneous region in an image. The threshold value is found such that variance in the region is minimized under change in the threshold value. Curve- and surface-fitting methods are also being developed that can accurately represent a region boundary in 2-D or 3-D with a parametric curve or a surface, respectively. A curve or a surface is optimized to minimize the number of control points representing a region with a prescribed accuracy. The optimized curve or surface is then revised by moving its control points interactively. Once a curve or a surface is found to accurately enclose a region of interest, it is quantized to produce the final 2-D region contour or 3-D region surface. These interactive tools can be used to revise unsatisfactory results obtained from any automatic segmentation method.


1. Morphing a sphere to a digital shape


2. Preliminary results

    Liver
    Human liver in digital form. (a)-(f) Sucessive approximations by polyhedron subdivision and corresponding mean errors. Final approximation contains 4,098 vertices and 8,192 triangular faces. (Dataset courtesy of University of Iowa, Dr. Sonka).
    (a) E=17.55
    (b) E=8.16
    (c) E=3.79
    (d) 1.95
    (e) E=1.20
    (f) E=0.83

    Liver tumor
    Liver tumor in digital form. (a)-(e) Sucessive approximations by polyhdron subdivision and corresponding mean errors. (f) Final approximation contains 4,098 vertices and 8,192 triangular faces. (Dataset courtesy of University of Iowa, Dr. Sonka).
    (a) E=8.54
    (b) E=3.68
    (c) E=1.67
    (d) 0.96
    (e) E=0.65
    (f)

    Left ventricular blood pool
    Left ventricular blood pool in digital form. (a)-(e) Sucessive approximations by polyhedron subdivision and corresponding mean errors. (f) Final approximation contains 1,026 vertices and 2,048 triangular faces. (Dataset courtesy of Rush Medical Center, Dr. Turner).
    (a) E=7.04
    (b) E=2.72
    (c) E=1.42
    (d) 0.79
    (e) E=0.49
    (f)

    Femoral stem (I)
    Femoral stem in digital form. (a)-(f) Sucessive approximations by polyhedron subdivision and corresponding mean errors. Final approximation contains 611 vertices and 1,218 triangular faces. (Dataset courtesy of the Miami Valley Hospital, Dr. Hangartner).
    (a) E=10.55
    (b) E=4.09
    (c) E=1.77
    (d) E=1.05
    (e) F=0.85
    (f) E=0.83

    Femoral stem (II)
    Femoral stem in digital form. (a)-(f) Sucessive approximations by polyhdron subdivision and corresponding mean errors. Final approximation contains 4,098 vertices and 8,192 triangular faces. (Dataset courtesy of University of Illinois, Chicago Medical Center, Dr. Barnald)
    (a) E=22.11
    (b) E=7.60
    (c) E=2.96
    (d) 1.48
    (e) E=0.92
    (f) E=0.58


[WSU Home Page] [CSE Department Home Page] [Intelligent Systems Lab]

For more information contact A. Goshtasby (mailto:ardy@cs.wright.edu).

Last modified:1/8/2001.