Date of Award

Summer 1986

Document Type

Thesis - Restricted

Degree Name

Master of Science (MS)

Department

Biomedical Engineering

First Advisor

Doerr, Thomas

Second Advisor

Hock, Jeffrey L.

Abstract

In the medical imaging field, x-ray computerized tomographic (CT) images provide views of thin cross-sections of the human anatomy. Typically, a series of CT images is obtained by scanning contiguous parallel cross-sections of a portion of a patient's body. The series of CT images represents a volume which may contain several components such as bone structures and soft-tissue organs. The components of the volume are referred to as objects. Traditionally, the images are viewed one at a time. The three-dimensional characteristics of a particular object can be inferred by mentally connecting the portions of the cross-sections where the object is visually identified. A surface of an object is the boundary that separates the object of interest from the surrounding objects. An alternative to mentally visualizing an object, consists of synthesizing an image of the surface of the object from the CT images. The algorithms used to synthesize images are known as 3D algorithms. In general, 3D algorithms create a synthesized image in two steps. The first step consists of extracting a mathematical representation of the object surface from the CT images. The second step consists of synthesizing an image from the mathematical representation of the object surface. The synthesized images are known as 3D images. The goal of the 3D algorithms is for the 3D images to look like a photograph of the object under examination. It has been shown that 3D images are useful in surgical procedures which require a direct visualization of the patient's anatomy. Craniofacial surgery is an example of a surgical procedure that benefits from the use of 3D images. 3D algorithms are computationally intensive. Software implementations of 3D algorithms have been demonstrated by several researchers. An example of a 3D algorithm is the cuberille algorithm. The cuberille algorithm has been implemented using the standard computer equipment supplied with the CT scanners. The limited computer power of the supplied computers forces a trade-off between the quality of the 3D images and the amount of time required to synthesize the images. This thesis describes an implementation of the cuberille algorithm on a large minicomputer. The goal of the research was to explore whether or not enhancements to the cuberille algorithm could be developed that generate 3D images of higher quality. It will be shown that enhancements to both the surface extraction and the surface display steps resulted in 3D images of higher quality.

Download

Share

COinS

Restricted Access Item

Having trouble?