Jesung Kim

Date of Award

5-2002

Document Type

Thesis

Degree Name

Master of Science

Department

Department of Electrical and Computer Engineering

First Advisor

Steven C. Gustafson, PhD

Abstract

In mammography, X-ray radiation is used in sufficient doses to be captured on film for cancer diagnosis. A problem lies in the inherent nature of X-rays to cause cancer. The resolution of the images obtained on film is directly related to the radiation dosage. Thus, a trade-off between image quality and radiation exposure is necessary to ensure proper diagnosis without causing cancer. A possible solution is to decrease the dosage of radiation and improve the image quality of mammograms using post- processing methods applied to digitized film images. Image processing techniques that may improve the resolution of images captured at lower doses include crispening, denoising, histogram equalization, and pattern recognition methods. The Wright Patterson Air Force Base Hospital Radiology Department sponsored this research and provided digitized images of the American College of Radiology (ACR) phantom, which is a model for mammogram image quality and classification. Side by side comparisons were performed of high dose images and low-dose images post-processed using the methods mentioned. The result was improved- resolution on mammography images for lower radiation doses. Thus, this research represents progress towards solving a problem that currently plagues mammography: exposure of patients to high doses of cancer- causing radiation to obtain quality mammography images. By improving the image quality of mammography images at lower radiation doses, the problem of cancer induced by high radiation exposure is alleviated.

AFIT Designator

AFIT-GE-ENG-02-35

DTIC Accession Number

ADA409113

Download

Share

COinS