Date of Award

Winter 1-31-1994

Document Type

Thesis

Degree Name

Master of Science in Biomedical Engineering - (M.S.)

Department

Biomedical Engineering Committee

First Advisor

Yun Q. Shi

Second Advisor

David S. Kristol

Third Advisor

L. Horn

Abstract

This thesis reports the research results on automatic contour extraction from high voltage electron microscope (HVEM) image of thick cross section montages of small blood vessels. The previous work on this subject, which was based on the conventional edge detection operations combined with edge linking, has proven inadequate to describe the inner structural compartments of microvessels. In this thesis, an active contour model (commonly referred to as "Snakes") has been applied to advance the previous work. Active contour models have proven themselves to be a powerful and flexible paradigm for many problems in image understanding, especially in contour extraction from medical images. With the developed energy functions, the active contour is attracted towards the edges under the action of internal forces (describing some elasticity properties of the contour), image forces and external forces by means of minimization of the energy functions. Based on this active model, an effective algorithm is implemented as a powerful tool for 2-D contour extraction in our problem for the first time. The results thus obtained turn out to be encouraging.

Share

COinS