Project Title:
Adaptive Wavelet Image Processing
05.01-2577
912031
Adaptive Wavelet Image Processing
Fastman, Inc.
1414 Millard Street
Bethlehem
PA
18018
Michael
Tucker
215-691-2577
MSFC
NAS8-39316
078
05.01-2577
912031
Abstract:
Adaptive, Wavelet Image Processing
The firm has recently developed the adaptive wavelet transform (AWT) and has shown
that the AWT is very adept at extracting features from complex one-dimensional signals
even in the presence of noise. This project will determine the ability of the AWT
to extract features from images in the presence of noise and variable lighting conditions.
Phase I will assess the feasibility of utilizing the AWT for robust image feature
detection by extending the current one-dimensional AWT software code to two dimensions;
by demonstrating the ability of the AWT to decompose images into basis elements highly
related to the images' features; by determining how to combine currently used pattern
classification and neural networks with an adaptive wavelet preprocessor to detect
object regions; and by investigating real-time operation of the AWT.
This research will lead to advances in real-time spectral analysis for many commercial
and NASA applications, including image processing and speech processing monitoring
of machine vibration signals. It can also provide high-ratio data compression for
telephone answering machines, voice mail systems, cellular phones, fax machines,
teleconferencing equipment, and other equipment.
wavelet transform, image processing, feature detection