Date of Award

Spring 1991

Document Type

Thesis - Restricted

Degree Name

Master of Science (MS)

Department

Electrical and Computer Engineering

First Advisor

Mulligan, Michael G.

Abstract

A new generation of x-ray based computed tomography (CT) medical imaging scanners is emerging which will provide high resolution images of human anatomy with one second scan times. The high speed data communication channel which carries x-ray detector information passes through an electrically and mechanically hostile environment known as the raw image channel. Data errors induced on this communication channel may cause visible image artifacts when the corrupted detector information is processed to-produce an image. In this thesis, the hardware implementation for real-time error control coding of x-ray detector data on the CT raw image channel is studied. To understand what impact corruption of data on the raw image channel has on the image quality of a reconstructed CT image, software simulations of raw data corruption are performed on a commercially available GE 9800 CT system. Once the corrupted raw data is reconstructed into a CT image, the resulting image artifacts are qualitatively analyzed. A Reed-Solomon block code is then combined with a Group-Coded Recording code to provide a concatenated coding scheme which is superior to other error control techniques for the given channel characteristics, since it is highly effective in correcting the burst errors which dominate the raw image channel. The Reed-Solomon and Group-Coded Recording codes are used extensively in magnetic and optical disk drives, and a strong correlation is developed between CT and disk drive architectures. After the concatenated code is fully defined and optimized to the unique characteristics of the raw image channel noise sources, modulation scheme, and system throughput requirements, practical implementation issues are introduced. These issues, which include decoder complexity reduction and hardware availability, lead to an overall coding system which substantially improves the reliability and image quality of a CT system and may be implemented with existing hardware technology.

Share

COinS

Restricted Access Item

Having trouble?