Signal modeling with non-uniform topology lattice filters

Abstract

This article presents a new class of constrained and specialized Auto-Regressive (AR) processes. They are derived from lattice filters where some reflection coefficients are forced to zero at a priori locations. Optimizing the filter topology allows to build parametric spectral models that have a greater number of poles than the number of parameters needed to describe their location. These NUT (Non-Uniform Topology) models are assessed by evaluating the reduction of modeling error with respect to conventional AR models.

Parametric Analysis of Ultrasound Backscatter Signals for Monitoring Cancer Cell Structural Changes During Cancer Treatment

“…Lattice filters are a well-known signal analysis and coding tool. Their parameters, the reflection coefficients, have a good robustness to noise and quantization effects [16]. In the lat tice formulation, the reflection coefficients can be computed by minimizing the norm of the forward residual or the back ward residual, or a combination of the two.…”

Parametric Analysis of Ultrasound Backscatter Signals for Monitoring Cancer Cell Structural Changes During Cancer Treatment

“…Lattice filters are a well-known signal analysis and coding tool. Their parameters, the reflection coefficients, have a good robustness to noise and quantization effects [16]. In the lat tice formulation, the reflection coefficients can be computed by minimizing the norm of the forward residual or the back ward residual, or a combination of the two.…”

Spectral Estimation Using Constrained Autoregressive (CAR) Model

“…Although there is some work on evaluation of the models in noisy conditions. Sacha has shown in his work [4] that the probability of error in resolving the two closely spaced peaks in PSD by using the AR model increases with decreasing SNR of the signal. Zhang gives [3] some idea about how a model can be changed by explicitly constraining any parameter value.…”