Overcomplete Expansions and Robustness

TY - GEN

T1 - Overcomplete Expansions and Robustness

AU - Cvetkovic, Zoran

AU - Vetterli, Martin

PY - 1998/12

Y1 - 1998/12

N2 - Manifestations of robustness of overcomplete expansions, acquired as a result of the redundancy, have been observed for two primary sources of degradation: white additive noise and quantization. These two casesrequire different treatments since quantization error has a certain structure which is obscured by statistical analysis. The central issue of this chapter is a deterministic analysis of quantization error of overcomplete expansions with a focus on Weyl-Heisenberg frames in L2(ℝ). This analysis demonstrates that overcomplete expansions exhibit a higher degree of robustness to quantization error than to a white additive noise, and indicates how this additional robustness can be exploited for improvement of representation accuracy.

AB - Manifestations of robustness of overcomplete expansions, acquired as a result of the redundancy, have been observed for two primary sources of degradation: white additive noise and quantization. These two casesrequire different treatments since quantization error has a certain structure which is obscured by statistical analysis. The central issue of this chapter is a deterministic analysis of quantization error of overcomplete expansions with a focus on Weyl-Heisenberg frames in L2(ℝ). This analysis demonstrates that overcomplete expansions exhibit a higher degree of robustness to quantization error than to a white additive noise, and indicates how this additional robustness can be exploited for improvement of representation accuracy.

U2 - 10.1016/S1874-608X(98)80012-5

DO - 10.1016/S1874-608X(98)80012-5

M3 - Other contribution

VL - 7

T3 - Wavelet Analysis and Its Applications: Signal and Image Representation in Combined Spaces

PB - Academic Press

ER -