- Start Date:
July 3, 2002
- Event Start Time:
12:30 PM
- Event End Time:
1:30 PM
- Organizers:
Jessica L. Herold
- Seminar Series:
REU Seminar
- Presenter(s):
Eva Curry - Rutgers University
- Event Location:
DIMACS Seminar room
- Abstract:
If we take a wavelet j(x), its translations by integers j(x-k) and the dyadic dilations j( 2j x-k) of all these functions, then we get a collection of functions which acts more or less like a basis for other functions. Writing a functions, such as sound or other signal or a line of pixels in an image, in terms of this “basis” is the first step in a lot of useful data analysis. For example, to filter noise in a signal you can remove terms corresponding to the noise in a wavelet expansion of the signal. Smoothing a blocky image can also be accomplished by changing only certain terms in the corresponding wavelet expansion. Both cases are examples of applying a filter to data. In this talk I will introduce multiresolution analysis wavelets (the standard, nice sort of wavelets) and what a filter is in this mathematical context. I will then describe my thesis project, which involves the classification of low-pass filters.
