Biomedical Engineering Reference
In-Depth Information
Segmentation has also been addressed with neural
networks in several applications
[2, 28, 37, 40, 69, 82,
101, 132]
.
6.4.8 Concluding remarks
Segmentation is an important step in many medical ap-
plications involving measurements, 3D visualization,
registration, and computer-aided diagnosis. This chapter
was a brief introduction to the fundamental concepts of
segmentation and methods that are commonly used.
Selection of the ''correct'' technique for a given ap-
plication is a difficult task. Careful definition of the goals
of segmentation is a must. In many cases, a combination
of several techniques may be necessary to obtain the
segmentation goal. Very often integration of information
frommany images (acquired from different modalities or
over time) helps to segment structures that otherwise
could not be detected on single images.
As new and more sophisticated techniques are being
developed, there is a need for objective evaluation and
quantitative testing procedures
[17, 20, 26]
. Evaluation
of segmentation algorithms using standardized protocols
will be useful for selection of methods for a particular
clinical application.
Clinical acceptance of segmentation techniques de-
pends also on ease of computation and limited user su-
pervision. With the continued increases in computer
power, the automated realtime segmentation of multi-
spectral and multidimensional images will become
a common tool in clinical applications.
Figure 6.4-11 Rendering of 3D anatomical models and 2D MRI
cross-sections of a patient with a meningioma. The models of the
skin surface, the brain, and the tumor (green) are based on
automatically segmented 3D MRI data. The precentral gyrus
(yellow) and the corticospinal tract (blue) are based on a previously
aligned digital brain atlas
[61]
. (Courtesy of Drs. Ron Kikinis, Mi-
chael Kaus, and Simon Warfield, Surgical Planning Lab, De-
partment of Radiology, Brigham and Women's Hospital, Boston.)
segmentation was carried out with the algorithm of Kaus
et al.
[60]
. This visualization was used to support pre-
operative surgical planning for tumor resection.
In some medical images, regions that have similar
average intensities are visually distinguishable because
they have different textures. Each pixel can be assigned
a texture value and the image can be segmented using
texture instead of intensity
[6, 79]
.
References
1. R. Adams, L. Bischof, ''Seeded region
growing,''
IEEE Trans. Pattern
Recogn. Mach. Intell.
, Vol. 16, no. 6,
pp. 641-647, 1994.
2. S. C. Amartur, D. Piraino, Y. Takefuji,
''Optimization neural networks for
the segmentation of Magnetic
Resonance images,''
IEEE Trans. Med.
Imag.
, vol. 11, no. 2, pp. 215-220,
1992.
3. M. S. Atkins, B. T. Mackiewich,
''Fully automatic segmentation of the
brain in MRI,''
IEEE Trans. Med.
Imag.
, vol. 17, no. 1, pp. 98-107,
1998.
4. G. J. Awcock, R. Thomas,
Applied
Image Processing.
NewYork:
McGraw-Hill, Inc., 1996.
5. R. Bajcsy, S. Kovacic,
''Multiresolution elastic matching,''
Comp. Vision, Graphic and Image
Proc.
, vol. 46, pp. 1-21, 1989.
6. D. G. Ballard, C. M. Brown,
Computer Vision
, Englewood Cliffs:
Prentice Hall, 1982.
7. P. A. Bandettini, A. Jesmanowicz,
E. C. Wong, J. S. Hyde, ''Processing
strategies for functional MRI of the
human brain,''
Magn. Res. Med.
, vol.
30, pp. 161-173, 1993.
8. I.N. Bankman, T. Nizialek, I. Simon,
O.B. Gatewood, I.N. Weinberg, W R.
Brody, ''Segmentation algorithms for
detecting microcalcifications in
mammograms'',
IEEE Trans. Inform.
Techn. Biomed.
, vol. 1, no. 2,
pp. 141-149.
9. J. Bernsen, ''Dynamic thresholding of
gray-level images,''in Proc.
8th Int.
Conf. Pattern Recognition
, Paris,
France, pp. 1251-55, Oct. 1986.
10. S. Beucher, ''Segmentation tools in
mathematical morphology,''
SPIE
,
vol. 1350, Image Algebra and
Morphological Image Processing,
pp. 70-84, 1990.
11. J.C. Bezdek, L.O. Hall, L.P. Clarke,
''Review of MR image segmentation
techniques using pattern
recognition,''
Med. Phys
., vol. 20,
no. 4, pp. 1033-1048, 1993.
12. M. Bomans, K.-H. Hohne, U. Tiede,
M. Riemer, ''3-D segmentation of MR
images of the head for 3-D
display,''
IEEE Trans. Medical
Imaging
, vol. 9, no. 2, pp. 177-183,
1990.
13. M.E. Bramdt, T.P. Bohan, L.A.
Kramer, J.M. Fletcher, ''Estimation of
CSF, white and gray matter volumes
in hydrocephalic children using fuzzy
clustering of MR imaging,''
