Narrow Results

This paper presents a data estimation scheme for wide band multichannel charge sampling filter bank receivers together with a complete system calibration algorithm based on the least mean squared (LMS) algorithm. A unified model has been defined for the receiver containing all first order mismatches, offsets, imperfections, and the LMS algorithm is employed to track these errors. The performance of this technique under noisy channel conditions has been verified. Moreover, a detailed complexity analysis of the calibration algorithm is provided which shows that sinc filter banks have much lower complexity than traditional continuous-time filter banks.

A novel digital compensation scheme for measuring, estimating and correcting linear weakly time-varying analog errors in frequency-interleaved analog-to-digital converters (FI-ADCs) is presented. This method features three important improvements over existing approaches: First, the Wigner–Ville distribution (WVD) is used to better estimate the nonstationary analog channel frequency response (ACFR) spectrum. Secondly, the estimated ACFR spectrum is approximated with a rational polynomial model using the ${\ell }_1$-norm metric. The corresponding ${\ell }_1$-norm minimization problem is solved using a primal-relaxed dual global optimization (PRD-GOP) method. Thirdly, the digital compensation circuitry is designed utilizing a preconditioned biconjugate gradient stabilized (BICGSTAB) algorithm that yields a computationally efficient solution. Numerical experimentations have been conducted and the outcomes validate the feasibility and superior performance of this proposed method.

A decomposition of signal into a set of frequency channels of equal bandwidth on a logarithmic scale, i.e., an analysis of the signal using constant Q filters, using wavelet and multiresolution analysis is used in this paper to derive the cepstral features for separated spatial frequency bands. Not like filter banking analysis, wavelet analysis decomposes signals into orthogonal spatial frequency bands, i.e., the overlap between two neighbor frequency bands is very small. Based on this property, channel weight can definitely be set to each frequency channel to increase the discriminability to distinguish between two signals. The recognition rate can then be improved. We use a Bayesian network to model each channel and propose an algorithm to give the channel weights. The experimental result shows that using 3-channel decompositions can get a better recognition rate than 1-channel recognition of the speech signals. The average recognition rate is also more superior than the filter-banking method and MFCC method by 3.54% and 1.95% respectively.

The matched wavelet is designed in this paper using an improved genetic algorithm for detecting the Heart Rate Variability (HRV) variations within phases of the menstrual cycle accurately. The idea of an improved genetic algorithm is to use an optimization technique like least mean square (LMS) before the genetic algorithm. The advantage of using the LMS prior to the genetic algorithm is to optimize the data before giving to the genetic algorithm, thereby limiting the area of the search for an optimal solution. The results show that matched wavelets created using an improved genetic algorithm can detect the HRV variations accurately in the standing and laying postures.

This paper describes an efficient and adaptive method of threshold estimation for removing impulse noise from images, based on Double Density Wavelet Transform (DDWT). The performance of image de-noising algorithms using wavelet transforms can be improved significantly by fixing an optimum threshold value, based on the analysis of the statistical parameters of subband coefficients. In this proposed method, the choice of the threshold estimation is carried out by analyzing the statistical parameters of the wavelet subband coefficients like standard deviation, arithmetic mean and geometrical mean. Here the noisy image is first decomposed into many levels to obtain different frequency bands using DDWT. Then soft thresholding method is used to remove the noisy coefficients, by fixing the optimum threshold value by the proposed method. Experimental results on several test images by using the proposed method show that, the proposed method yields significantly superior image quality and better Peak Signal-to-Noise Ratio (PSNR). Some comparisons with the best available results will be given in order to illustrate the effectiveness of the proposed algorithm.

In this paper, an off-line double density discrete wavelet transform based de-noising and baseline wandering removal methods are proposed. Different levels decomposition is used depending upon the noise level, so as to give a better result. When the noise level is low, three levels decomposition is used. When the noise level is medium, four levels decomposition is used. When the noise level is high, five levels decomposition is used. Soft threshold technique is applied to each set of wavelet detail coefficients with different noise level. Donoho's estimator is used as a threshold for each set of wavelet detail coefficients. The results are compared with other classical filters and improvement of signal to noise ratio is discussed. Using the proposed method the output signal to noise ratio is 19.7628 dB for an input signal to noise ratio of -7.11 dB. This is much higher than other methods available in the literature. Baseline wandering removal is done by using double density discrete wavelet approximation coefficients of the whole signal. This is an unsupervised method allowing the process to be used in off-line automatic analysis of electrocardiogram. The results are more accurate than other methods with less effort.