Novel vector rank M-type K-nearest neighbor filters are adapted for multichannel image processing. Ultrasound images are contaminated by impulsive noise and 3-D filtering algorithms are applied.
Novel vector rank M-type K-nearest neighbor filters are adapted for multichannel image processing. Ultrasound images are contaminated by impulsive noise and 3-D filtering algorithms are applied.
A research analysis on item-based algorithms for collaborative filtering is presented. The aim of the presented activity was to find a configuration of an item-based algorithm capable of providing good results but als...
详细信息
A research analysis on item-based algorithms for collaborative filtering is presented. The aim of the presented activity was to find a configuration of an item-based algorithm capable of providing good results but also independent from the data set. Four data sets were used for the algorithm validation: Netflix, MovieLens, BookCrossing, and Jester. The experimentation involved the following aspects: similarity computation, size of the neighbourhood, prediction computation, minimum number of co-rated items. Results were evaluated in terms of root mean squared error (RMSE). The result of the activity is an independent domain configuration for an item-based algorithm which produced satisfactory results with most of the above mentioned data sets.
In many applications it is desirable to use filters with linear phase characteristics. This paper presents an approach for developing such filters for adaptive signal processing. Several least-squares algorithms are d...
详细信息
In many applications it is desirable to use filters with linear phase characteristics. This paper presents an approach for developing such filters for adaptive signal processing. Several least-squares algorithms are derived for adjusting the coefficients of an adaptive linear phase filter. It is shown that smoothing filters, rather than the commonly used prediction filters, are a more natural choice if linear phase characteristics are required.
The authors present an approach to the development of fast and numerically stable recursive least squares (RLS) algorithms for adaptive nonlinear filtering using QR-decomposition of the data matrix. They introduce a p...
详细信息
The authors present an approach to the development of fast and numerically stable recursive least squares (RLS) algorithms for adaptive nonlinear filtering using QR-decomposition of the data matrix. They introduce a pair of QR-RLS adaptive algorithms for second-order Volterra filtering. Both the algorithms are based solely on Given's rotation. Hence both are numerically stable and highly amenable to parallel implementations using arrays. One of the algorithms is a block processing algorithm in the sense that it processes all the channels simultaneously. The other processes the channels sequentially. The sequential algorithm is computationally much more efficient than the block algorithm and is comparable to that of fast RLS Volterra filters. Another attractive feature of sequential processing is that knowledge of the single-channel algorithm can be applied to the multichannel case.< >
Stochastic gradient algorithms are widely used in signal processing. Whereas stopping rules for deterministic descent algorithms can easily be constructed, using for instance the norm of the gradient of the objective ...
详细信息
Stochastic gradient algorithms are widely used in signal processing. Whereas stopping rules for deterministic descent algorithms can easily be constructed, using for instance the norm of the gradient of the objective function, the situation is more complicated for stochastic methods since the gradient needs first to be estimated. We show how a simple Kalman filter can be used to estimate the gradient, with some associated confidence, and thus construct a stopping rule for the algorithm. The construction is illustrated by a simple example. The filter might also be used to estimate the Hessian, which would open the way to a possible acceleration of the algorithm. Such developments are briefly discussed.
We describe and analyze two approaches to the implementation of the Conjugate Gradient algorithm for adaptive filtering. In particular their convergence rate and misadjustment are compared. A new analysis approach in ...
详细信息
We describe and analyze two approaches to the implementation of the Conjugate Gradient algorithm for adaptive filtering. In particular their convergence rate and misadjustment are compared. A new analysis approach in the z-domain is used in order to find the asymptotic performance, and stability bounds are established. The behavior of the algorithms in finite word-length computation are described and dynamic range considerations are discussed. It is shown that, close to steady-state, the algorithms' behaviors are similar to the Steepest Descent algorithm, where the stalling phenomenon is also observed. Using 16-bit fixed-point number representation, our simulations show that the algorithms are numerically stable.
A new approach to moving-average filtering is proposed based on a model of an input mixture as a finite sum of sinusoidal segments. The correspondent filter is particularly efficient when processing very short-duratio...
详细信息
A new approach to moving-average filtering is proposed based on a model of an input mixture as a finite sum of sinusoidal segments. The correspondent filter is particularly efficient when processing very short-duration mixtures. Among its advantages are high precision of useful signal recovery and relative simplicity of construction. algorithms and the results of computer simulations of the separation of signals as well as of the filtering of signals from interference including white noise are presented.< >
In this paper we present a family of "new" two-dimensional adaptive filtering algorithms for image processing applications. These algorithms are multidimensional versions of the families of data-reusing and ...
详细信息
In this paper we present a family of "new" two-dimensional adaptive filtering algorithms for image processing applications. These algorithms are multidimensional versions of the families of data-reusing and projection algorithms. These two classes of algorithms allow the adaptive filtering system designer to choose performance and computational complexity by changing parameters without actually changing algorithm structure. By changing parameters, the desired convergence rate can be achieved at the expense of additional computational complexity. Experiments show that significant improvement may be obtained by marginal increases in computational complexity over the traditional normalized LMS algorithm.
This paper proposes a new family of approximate QR-based least squares (LS) adaptive filtering algorithms called p-TA-QR-LS algorithms. It extends the TA-QR-LS algorithm by retaining different number of diagonal plus ...
详细信息
This paper proposes a new family of approximate QR-based least squares (LS) adaptive filtering algorithms called p-TA-QR-LS algorithms. It extends the TA-QR-LS algorithm by retaining different number of diagonal plus off-diagonals (denoted by an integer p) of the triangular factor of the augmented data matrix. For p=1 and N it reduces respectively to the TA-QR-LS and the QR-RLS algorithms. It not only provides a link between the QR-LMS-type and the QR-RLS algorithms through a well-structured family of algorithms, but also offers flexible complexity-performance tradeoffs in practical implementation. These results are verified by computer simulation and the mean convergence of the algorithms is also analyzed
As one of the most successful recommender systems, collaborative filtering (CF) algorithms can deal with high sparsity and high requirement of scalability amongst other challenges. Bayesian belief nets (BNs), one of t...
详细信息
As one of the most successful recommender systems, collaborative filtering (CF) algorithms can deal with high sparsity and high requirement of scalability amongst other challenges. Bayesian belief nets (BNs), one of the most frequently used classifiers, can be used for CF tasks. Previous works of applying BNs to CF tasks were mainly focused on binary-class data, and used simple or basic Bayesian classifiers (Miyahara and Pazzani, 2002; Breese et al., 1998). In this work, we apply advanced BNs models to CF tasks instead of simple ones, and work on real-world multi-class CF data instead of synthetic binary-class data. Empirical results show that with their ability to deal with incomplete data, extended logistic regression on naive Bayes and tree augmented naive Bayes (NB-ELR and TAN-ELR) models (Greiner et al., 2005) consistently perform better than the state-of-the-art Pearson correlation-based CF algorithm. In addition, the ELR-optimized BNs CF models are robust in terms of the ability to make predictions, while the robustness of the Pearson correlation-based CF algorithm degrades as the sparseness of the data increases
暂无评论