An ef®cient implementation of the Volterra ®lter is presented which uses a frequency domain representation to reduce the number of computations. The multidimensional convolution of the Volterra ®lter is transformed to the frequency domain giving a transformed input matrix which is sparse and obtained directly from a one-dimensional Fourier transform. In addition to the sparse nature of the tran...

Missing data interpolation problems can be conveniently preconditioned by re-cursive inverse ltering. A helix transform allows us to implement this idea in the multidimensional case. We show with examples that helix preconditioning can give a magnitude-order speedup in comparison with the older methods.

We present a new semide nite programming approach to FIR lter design with arbitrary upper and lower bounds on the frequency response magnitude. It is shown that the constraints can be expressed as linear matrix inequalities (LMIs), and hence they can be easily handled by recent interior-point methods. Using this LMI formulation, we can cast several interesting lter design problems as convex or ...

In this paper the approximation of a complex-valued speci cation by the frequency response of a 2-D IIR separable-denominator (SD) digital lter is considered. The approximation problem is transformed into an equivalent one, where a real-valued 2-D IIR SD digital lter with some additional characteristics has to be determined that approximates a given real-valued 2-D FIR digital lter. A theorem i...

We present a new architecture to realize a fully programmable rank order lter (ROF), based on Capacitive Threshold Logic (CTL) gates. Variants of ROFs, especially median lters, are widely used in digital signal and image/video processing and image enhancement. The CTL realization of the majority gates used in the ROF architecture allows the lter rank and the window size to be user-programmable,...

{ The leaky LMS adaptive lter can be implemented either directly or by adding random white noise to the input signal of the LMS adaptive lter. In this paper, we analyze and compare the mean-square performances of these two adaptive lter implementations for system identiication tasks with zero mean i.i.d. input signals. Our results indicate that the performance of the direct implementation is su...

We discuss the advantages of using overdetermined mixtures to improve upon blind source separation algorithms that are designed to extract sound sources from acoustic mixtures. A study of the nature of room impulse responses helps us choose an adaptive lter architecture. We use ideal inverses of acquired room impulse responses to compare the eeectiveness of diierent-sized separating lter conngu...

We discuss the relation between lattice and ladder structures for two-channel lter banks. It is wellknown that both lattice and ladder steps are powerful enough to generate all perfect reconstructing lter banks provided that the lter coe cients may take arbitrary values in a eld. However, we will show that the two concept di er in general. We relate the two concepts by looking at three properti...

This paper investigates the realization of optimal lter bank-based cepstral parameters. The framework is the Discriminative Feature Extraction method (DFE) which iteratively estimates the lter-bank parameters according to the errors that the system makes. Various parameters of the lter-bank, such as center frequency, bandwidth, gain are optimized using a string-level optimization and a frame-le...

We present a new semideenite programming approach to FIR lter design with arbitrary upper and lower bounds on the frequency response magnitude. It is shown that the constraints can be expressed as linear matrix inequalities (LMIs), and hence they can be easily handled by recent interior-point methods. Using this LMI formulation, we can cast several interesting lter design problems as convex or ...