Cardiac Tracking in Echocardiac Sequences
- Introduction
- The challenges in echocardiographic tracking
- Research
- Related links
- Background reading
- Papers of Interest
Introduction
Two-Dimensional (2-D) image sequence data from most available cardiac
imaging modalities have been used in attempts to quantify (regional)
left ventricular (LV) (endocardial) motion.
Two-dimensional echocardiography is the preferred cardiac imaging method
for reasons as:
- At all levels of exposure and dosage currently in use,
ultrasound seems to be the safest diagnostic technique.
- Real-time ultrasonic imaging is an interactive process that
provides two-dimensional information about structure and
function, i.e. flexibility in image acquisition.
The challenges in echocardiographic tracking
There has been much recent work in the computer vision community on
tracking and understanding object motion. There are, however,
a number of significant problems with tracking cardiac motion,
and hence to derive clinical parameters based on wall motion:
- There are inherent problems of measuring the motion of a 3-D
spatially deforming object from 2-D images.
- The heart undergoes a nonrigid motion.
- There is a need for real-time analysis.
Echocardiographic image tracking is more challenging because:
- Ultrasound images have a low signal-to-noise ratio.
- Ultrasound does not pass through bone or air.
- Ultrasound images can contain significant imaging artifacts.
- Spatial resolution in ultrasound images is non-uniform.
Research
2-D
Snakes, or active contour models, have been used in much of the
work, which use physics based descriptions of the contours
to constrain the solution (Metaxas & Terzopoulos 1993;
Cohen, Ayache & Sulgerr 1992).
Sonka et al. (1995) apply dynamic programming
techniques to segment the heart borders.
Mailloux et al. (1987) measure wall motion
using optical flow.
The
Oxford group tracks heart motion using Blake's
contour tracker.
3-D
Feldmar & Bardinet at
INRIA track the heart using superquadrics.
Park et al. (1996) use deformable models on
tagged MRI data.
Related links
The following groups are also involved in cardiac motion estimation.
Background reading
- Feigenbaum, H. Echocardiography, 5th edition,
Philadelphia, PA: Lea & Febiger, 1994.
- Kremkau, F. W. Diagnostic Ultrasound: Principles and Instruments,
4th edition. W. B. Saunders Company, Philadelphia, PA, USA, 1993.
- Webb, S. The physics of medical imaging, IOP.
Papers of Interest
-
Ayache, N., Cohen, I. and Herlin, I.,
Medical image tracking. In Active Vision
(A. Blake and A. Yuille, Eds.) 285-302, MIT.
-
Blake, A., Isard, M. and Reynard, D.
Learning to track the visual motion of contours.
Art. Intell., 78, pp. 179-212, 1995.
-
Cootes, T. F., Taylor, C. J., Cooper, D. H and Graham, J.,
Active shape models - their training and application.
CVGIP,61, pp. 38-59, 1995.
-
Herlin, I. and Ayache, N.,
Feature extraction and analysis methods for
sequences of ultrasound images,
in Proceedings ECCV, pp. 43-57, 1992.
-
Evans, A. N. and Nixon. M. S.,
Biased motion-adaptive temporal filtering for speckle reduction
in echocardiography.
IEEE Transactions Med. Imag., 15:1, pp. 39-50, 1996.
-
McEachen, J. C. II and Duncan, J. S.,
Shape-based tracking of the left ventricular wall motion.
IEEE Trans. Medical Imaging, 16:3, pp.270-283, 1997.
-
Perez, J. E., Waggoner, A. D., Barzilai. B., Melton,. H. E., Miller,
J. G., and Sobel, B. E. On-line assessment of ventricular function
by automatic boundary detection and ultrasonic backscatter imaging,
Journal of the American college of Cardiology, 19, 313-320.
1992.
-
Wells, P. N. T.
Advances in Ultrasound Techniques and Instrumentation,
Churchill Livingstone, NY, 1993.