New Results in the Theory and Applications of Coiflet-Type Wavelets
Dong Wei
Laboratory for Vision Systems
Department of Electrical and Computer Engineering
The University of Texas at Austin, Austin, TX 78712-1084
Monday, March 3rd, 3:00pm -- 4:00pm, ENS 602
Fourier techniques are tools for analyzing the behavior of functions in the frequency domain, but Fourier techniques cannot be effectively used to isolate the occurrence of frequency characteristics in the function domain (``time''). Wavelet techniques, however, are well-suited at localizing phenomenon in the time and frequency. Despite its short history, wavelet theory has found applications in a remarkable diversity of disciplines such as mathematics, physics, numerical analysis, signal processing, probability theory and statistics, and computer graphics. In signal processing, wavelets are excellent representations of nonstationary signals.
After briefly introducing the fundamentals of wavelet and filter bank theory, I will describe two novel families of one-dimensional wavelets. These new wavelet families are promising in a wide range of digital signal processing applications, such as signal sampling, interpolation, and approximation, fast transforms, and image data compression. In addition, I will propose future research in extending these new wavelet families for non-separable multidimensional signal processing.
Supervisors: Prof. Alan C. Bovik and Prof. Brian L. Evans
A list of digital signal processing seminars is available at from the ECE department Web pages under "Seminars". The Web address for the digital signal processing seminars is http://www.ece.utexas.edu/~bevans/dsp_seminars.html