Master's Thesis Defense - A Practical Application for Noise Power Spectrum Analysis
Date: May 22, 2008
Time: 9:00 AM
Location: Bossone Research Enterprise Center, Room: 709
Scott L. Cupp
Advisors: Chang Chang, Ph.D. and Andrew D. Maidment, Ph.D.
Noise power spectrum (NPS) analysis is a useful image quality matrix. It provides a quantitative description of the amount and frequency of the noise fundamentally contained within a particular imaging system. The calculation of the noise power spectrum of an imaging system has been greatly facilitated recently with the use of modern mathematical programs, such as MATLAB®, which can quickly calculate the two dimensional Fast Fourier Transform (FFT) of an image. However an application that allows users to conveniently import image files, appropriately manipulate data and effortlessly generate two and one dimensional noise power spectrum in a standardized way remains lacking. The goal of this project is to develop a graphical user interface (GUI) equipped application, utilizing MATLAB® as the underlying program, which will allow imaging scientists, as well as clinical diagnostic imaging medical physicists, to quickly and easily generate NPS analysis. The resulting application allows the user to import a variety of common image formats, load single or multiple image realizations, linearize image data using common methods, graphically or numerically select a region of interest (ROI), select sub-ROI sampling area matrixes, calculate the NPS using two common zero frequency reduction methods, and generate one-dimensional NPS graphs along the different frequency axes. Further development of this program will allow it to be used as a standalone application without the user requiring access to the full version of MATLAB®.
The Bossone Research Enterprise Center is located at the corner of 32nd and Market Streets.