Enhanced resolution and modeling of exponential signals
- 1 January 1991
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 3033-3036 vol.5
- https://doi.org/10.1109/icassp.1991.150094
Abstract
A signal enhancement algorithm is described which achieves the objective of modification of the given data set that then serves as a cleansing process whereby corrupting noise, measurement distortion, or theoretical mismatch present in the given data set is removed. Particular attention is directed toward properties that are describable using a singular value decomposition of a data generated matrix. An example is given demonstrating a significant improvement in the performance of subspace-based frequency estimation techniques. The algorithm is shown to provide a useful means for solving a variety of important signal processing problems. It has been successfully applied to the missing data problem, deconvolution, high-dimensional filter synthesis, and image synthesis.Keywords
This publication has 3 references indexed in Scilit:
- Signal enhancement-a composite property mapping algorithmIEEE Transactions on Acoustics, Speech, and Signal Processing, 1988
- Image Restoration by the Method of Convex Projections: Part 1ߞTheoryIEEE Transactions on Medical Imaging, 1982
- Estimation of frequencies of multiple sinusoids: Making linear prediction perform like maximum likelihoodProceedings of the IEEE, 1982