Date of Award

Spring 1991

Document Type

Thesis - Restricted

Degree Name

Master of Science (MS)

Department

Biomedical Engineering

First Advisor

Myklebust, Joel

Second Advisor

Jeutter, Dean C.

Abstract

Electrical Stimulation of the spinal cord results in alterations in the metabolic activity represented by local glucose utilization rate. Studies have shown a close relationship between local functional activity and energy metabolism in the central nervous system. An advanced computerized method was developed that provides correct evaluation, detection, and visualization of such changes in an estimate of its magnitude, frequency and region of occurrence. X-ray autoradiographic images represented only relative concentrations of glucose isotope contained in deoxyglucose-6-phosphate in various tissues. Resolution of differences in relative rates is limited to human visualization of differences in shades of gray levels. This project was conducted by development of a three dimensional reconstruction method and various image processing techniques for evaluation of metabolic activity of the spine. The Projection of spinal images in three dimensions requires different stages of development. A preprocessing stage is concerned with the development and application of pattern recognition techniques for detection and removal of background noise in the digitized images. A backpropagation neural network model was developed application of pattern recognition techniques for detection and removal of background noise in the digitized images. A backpropagation neural network model was developed for local feature detection of the spinal orientation and removal of background. The reconstruction stage is divided into two major areas. First, development of alignment algorithms for presentation of accurate reconstruction. Two algorithms are developed: One algorithm accounts for principal axis alignment methods and the other accounts for alignment via cross correlation of two images. Both algorithms contribute for correction in rotational and translational alignment. Second, The reconstruction algorithm processes images by the pixel shift technique.So far the reconstruction process allows rotation of the body around the axis of reconstruction. Horizontal sectioning of the reconstructed body at various levels of the view field is also possible. Post processing methods include: Metabolic activity evaluation of glucose utilization via Sokoloff technique, pseudo coloring or density slicing of metabolic activity rates to colored images, and various frequency domain analysis methods: Cross-spectrum, Coherence, and Frequency Responses. ..

Download

Share

COinS

Restricted Access Item

Having trouble?