New Techniques in Signal and Image Processing
Abstract
A multidimensional periodicity detection algorithm is established, and implementation has been made on ECoG data and with broadband and I/f noise. The algorithm requires the characterization of periodic continuous wavelet transforms, and the solution of an optimization problem coupled with an averaging method for mixed-norm spaces. A transition of this technology has been made to Techno-Sciences, Inc. in their NASA supported work for Health Maintenance of jet engines using eddy current sensors. A multidimensional irregular sampling technique is also established, and the corresponding algorithm solves signal reconstruction problems in MRI for sampled spectral data on interleaving spirals related to fast imaging techniques. The tools required are our results from the theory of frames, as well as the Beurling-Landau analysis of balayage.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1999
- Accession Number
- ADA380026
Entities
People
- John Benedetto
Organizations
- University of Maryland