Segmentation of Angiographic Images

Radiographic imaging has proven to be an invaluable tool for the diagnosis and treatment of heart disease. To aid in this task, we are developing methods to perform the automated segmentation of coronary arterial structure from data contained in ciné sequences of biplanar x-ray angiograms. The overall goal is to detect, locate and quantitatively characterize the vasculature, especially stenotic lesions.

The images we are working with are acquired using biplanar imaging devices which take a combined total of 30 images per second. Due to radiographic constraints, the two views are interleaved and the result is a pair of 15 images per second ciné sequences as can be seen in the following mpeg sequences.

These images are the input used by the system we are developing. As mentioned above, the two views are taken simultaneously (+/- 1/15 sec). The upper and lower halves of the image correspond each to a view.

These next images are the result of processing with the first stage of our segmentation system. The results are obtained with virtually no user interaction.

The final stages in our system require user interaction. Through the use of the user interface, users may edit the initial segmentation created by the system. Because much of the segmentation has already been done, the time required of the user to finish the segmentation is minimal.

Project Members:

• Dr. Norberto F. Ezquerra
• James F. O'Brien

References:

• "Model-Based Labeling of Arterial Structure Using Temporal Disambiguation,"
  N. Ezquerra, S. Capell, L. Klein, and P. Duijves. submitted to IEEE Trans. on Medical Imaging.

• "Three-Dimensional Representation of Coronary Arteries: A Preliminary Clinical Evaluation,"
  L. Klein, J. Peifer, M. Ghazzal, N. Ezquerra, C. Cooke and E. Garcia. Circulation, Vol. 84, No. 4, II-721, 1995.

• "The Automated Segmentation of Coronary Vessels in Angiographic Image Sequences,"
  J. F. O'Brien and N. F. Ezquerra. IEEE Proc.: Visualization in Biomedical Computing, October 1994.

• "Automated Segmentation of Coronary Vessels in Angiographic Image Sequences Using Temporal, Spatial, and Structural Constraints,"
  J. O'Brien and N. Ezquerra. Visualization in Biomedical Computing (VBC T94) Conference, SPIE Vol. 2359, No. 25, pp. 25-37, October 1994.

• "Spatiotemporal Detection of Arterial Structures Using Active Contours,"
  E. Hyche, N. Ezquerra, and R. Mullick. Proc. 2nd. Int. Conf. on Visualization in Biomedical Computing (VBC T92), SPIE 1808, R. Robb, ed., pp. 56-62, Chapel Hill, NC, October 1992.

• "Three-Dimensional Coronary Angiography,"
  L. Klein, E. Garcia, C. Cooke, R. Folks and J. Peifer. Am. J. Cardiac Imaging Vol. 7, No. 3, pp. 187-194, 1992.

• "Coronary Vasculature Visualization from Limited Angiographic Views,"
  J. W. Peifer, R. Mullick, N. F. Ezquerra, M. E. Hyche, E. V. Garcia, L. Klein, and C. D. Cooke. IEEE Proc. of the First Conference on Visualization in Biomedical Computing, Atlanta, GA, pp. 195-200, May 1990.

• "3-Dimensional Reconstruction of Arterial Structure from Biplane Angiography,"
  C. D. Cooke, L. Jofre, L. Klein, J. W. Peifer, N. F. Ezquerra, W. S. Briggs, M. S. West, and E. V. Garcia. IEEE Technicon '87 Conference Proceedings, Miami, Florida, October 28-30, 1987.