In this paper, comparison of various orthogonal transforms in Wiener filtering is discussed. The study involves the family of discrete orthogonal transforms called Complex Hadamard Transform, which has been recently introduced by the same authors. Basic definitions, properties and transformation kernel of Complex Hadamard Transform are also shown.

This paper is concerned with blind identification of a nonminimum phase transfer function model. By over-sampling the output at a higher rate than the input, it is shown that its input-output relation can be described by a single input multiple output model (SIMO) with a common denominator polynomial. Based on the model expression, we present an algorithm to estimate numerator polynomials and common denominator polynomial in a blind manner. Furthermore, identifiability of the proposed scheme is clarified, and some numerical results are given for demonstrating its effectiveness.

This paper presents a new technique for the synthesis of sets of low-peak sequences exhibiting low peak cross correlation. The sequences also have flat power spectra and are suitable for many applications requiring such sets of uncorrelated pseudo-white-noise sources. This is a new application of the ta-sequence (trigonometric function aliasing sequence), which itself is a very new technique that uses the well-known "Reed-Solomon code" or "One coincident code" to generate these sets of low-peak-factor pseudo-white-noise exhibiting low peak cross correlation. The ta sequence method presented here provides the means for generating various sequences at the lengths required for such applications as system measurement (needing uncorrelated test signals), pseudo-noise synthesis (for spread spectrum communication), and audio signal processing for sound production (for enhancing spatial imagery in stereo signals synthesized from mono sources) and sound reproduction (for controlling unwanted interference effects in multiple-loudspeaker arrays).

Wavelet filters used in usual applications are not time-varying filters. In this paper, we present a novel method to design biorthogonal wavelet filters which are orthogonal to the input signals. We call newly designed filters time-varying lifting wavelet filters (TVLWF). Their feature is to vary the wavelet filters adapting to the input signal by tuning free parameters contained in the lifting scheme developed by Sweldens. These filters are almost compact support and perfect reconstruction. By using TVLWF, we demonstrate an application to data compression of electrocardiogram (ECG) which is one of the semi-periodic time-series signals and show that the time-varying system can be constructed easily and the proposed method is very useful for data compression.

A nonlinear time series predictor was proposed, in which a nonlinear sub-predictor (NSP) and a linear sub-predictor (LSP) are combined in a cascade form. This model is called "hybrid predictor" here. The nonlinearity analysis method of the input time series was also proposed to estimate the network size. We have considered the nonlinear prediction problem as a pattern mapping one. A multi-layer neural network, which consists of sigmoidal hidden neurons and a single linear output neuron, has been employed as a nonlinear sub-predictor. Since the NSP includes nonlinear functions, it can predict the nonlinearity of the input time series. However, the prediction is not complete in some cases. Therefore, the NSP prediction error is further compensated for by employing a linear sub-predictor after the NSP. In this paper, the prediction mechanism and a role of the NSP and the LSP are theoretically and experimentally analyzed. The role of the NSP is to predict the nonlinear and some part of the linear property of the time series. The LSP works to predict the NSP prediction error. Furthermore, predictability of the hybrid predictor for noisy time series is investigated. The sigmoidal functions used in the NSP can suppress the noise effects by using their saturation regions. Computer simulations, using several kinds of nonlinear time series and other conventional predictor models, are demonstrated. The theoretical analysis of the predictor mechanism is confirmed through these simulations. Furthermore, predictability is improved by slightly expanding or shifting the input potential of the hidden neurons toward the saturation regions in the learning process.

In this paper we propose a new pipelined architecture for the LMS adaptive filter which can be implemented with less than half the amount of calculation needed for the conventional architectures. Even though the proposed architecture reduces the required calculation, it can simultaneously produce good convergence characteristics, a short latency and high throughput characteristics.

This paper proposes a fast encoding algorithm for iterated function system (IFS) coding of gray-level homogeneous fractal images. In order to realize IFS coding of high order fractal images, it is necessary to solve a set of simultaneous equations with many unknowns. Solving the simultaneous equations using a multi-dimensional, numerical root-finding method is however very time consuming. As preprocessing of numerical computation, the proposed algorithm employs univariate polynomial manipulation, which requires less computation time than multivariate polynomial manipulation. Moreover, the symmetry of the simultaneous equations with respect to the displacement coefficients enables us to derive an equation with a single unknown from the simultaneous equations using univariate polynomial manipulation. An experimental result is presented to illustrate that the encoding time of the proposed algorithm is about 5 seconds on a personal computer with a 400 MHz Pentium II processor.

Recently, Kundur and Hatzinakos showed that a linear restoration filter designed by using the almost obvious a priori knowledge on the original image, such as (i) nonnegativity of the true image and (ii) the smallest rectangle encompassing the original object, can realize a remarkable performance for a blind image deconvolution problem. In this paper, we propose a new set-theoretic blind image deconvolution scheme based on a recently developed convex projection technique called Hybrid Steepest Descent Method (HSDM), where some partial information can be utilized set-theoretically by parallel projections onto convex sets while the others are incorporated in a cost function to be minimized by a steepest descent method. Numerical comparisons with the standard set-theoretic scheme based on POCS illustrate the effectiveness of the proposed scheme.

A problem in image recognition in practical circumstances is that an observed image is often degraded by an imaging system. A conventional method in such a case is first to estimate the parameters of the imaging system and then restore the image before analysis. Here, we propose an alternative approach based on phase invariants in Fourier domain that needs no restoration and is fairly robust against both blur and noise. We show that the image phases in positive region of the Fourier transform of the point spread function (PSF) are blur-invariant provided that the PSF is central symmetric. Under the phase-invariant assumption, a phase correlation function between a standard image and the degraded image is used in developing the recognition algorithm. The effectiveness of this algorithm is demonstrated through experiments using ten classes of figure images from car license plates.

In this paper, a new scheme for designing multilevel BTC coding is proposed. Optimal quantization can be obtained by selecting the quantization threshold with an exhaustive search. However, this requires an enormous amount of computation and is, thus impractical when we consider an exhaustive search for the multilevel BTC. In order to find a better threshold so that the average mean square error between the original and reconstructed images is a minimum, the genetic algorithm is applied. Comparison of the results of the proposed method with the exhaustive search reveal that the former method can almost achieve optimal quantization with much less computation than that required in the latter case.

In this paper, a new feature which is characterized by the extrema density of 2-D wavelet frames estimated at the output of the corresponding filter bank is proposed for texture segmentation. With and without feature selection, the discrimination ability of features based on pyramidal and tree-structured decompositions are comparatively studied using the extrema density, energy, and entropy as features, respectively. These comparisons are demonstrated with separable and non-separable wavelets. With the three-, four-, and five-category textured images from Brodatz album, it is observed that most performances with feature selection improve significantly than those without feature selection. In addition, the experimental results show that the extrema density-based measure performs best among the three types of features investigated. A Min-Min method based on genetic algorithms, which is a novel approach with the spatial separation criterion (SPC) as the evaluation function is presented to evaluate the segmentation performance of each subset of selected features. In this work, the SPC is defined as the Euclidean distance within class divided by the Euclidean distance between classes in the spatial domain. It is shown that with feature selection the tree-structured wavelet decomposition based on non-separable wavelet frames has better performances than the tree-structured wavelet decomposition based on separable wavelet frames and pyramidal decomposition based on separable and non-separable wavelet frames in the experiments. Finally, we compare to the segmentation results evaluated with the templates of the textured images and verify the effectiveness of the proposed criterion. Moreover, it is proved that the discriminatory characteristics of features do spread over all subbands from the feature selection vector.

In this paper, we present a novel architecture named as Window-MSPA architecture which targets to window operations in image processing. We have previously developed a Memory Sharing Processor Array (MSPA) for fast array processing with regular iterative algorithms. Window-MSPA tries to optimize the data I/O ports and the number of processing elements so as to reduce hardware cost. The input scheme of image data is restricted to row by row input which simplifies the I/O architecture. Under this practical I/O restriction, the fastest processings are achieved. In this paper, we present the general Window-MSPA design methodology for wide variety of applications. As an practical application, we have already reported the design of MP@HL MPEG2 Motion Estimator LSI. Design formulas for Window-MSPA architecture are given for various size of window operations in image processing. Thus, the derived architecture is flexible enough to satisfy user's requirement for either area or speed.

In many applications of digital video database systems such as digital library, video data is often compressed with MPEG video algorithms. It will be an important technique to insert the additional information data like indexes and contents effectively into video database which is compressed with MPEG, because we can always deal with the additional information with video data itself easily. We propose a method for inserting optional binary data such as index information of digital library into MPEG-1 and -2 bitstreams. The binary data inserted MPEG video bitstreams using our proposed scheme are also according to the specification of the MPEG video frame structure. The proposed method allows us to extract the inserted binary data perfectly though MPEG-1 and -2 video are lossy algorithms. And the quality of decoded images after extracting added information is almost the same as that of ordinary MPEG bitstreams. Furthermore, traditional standard MPEG-1 and -2 video decoder which can not extract inserted binary data can also decode images from the binary data inserted MPEG video bitstreams without obvious image degradation. There are some different points between the proposed insertion technique of the binary data and the watermarking technique. The technique of watermarking prepares to deal with alter watermarking by others. And the technique of watermarking is required for the identification of the signature and the perfect extraction of the inserted image signature is not required in the lossy MPEG video environment. On the other hand, we have to extract all of the inserted binary information data correctly with the insertion technique of the binary information. Simulations using MPEG video sequences with inserted binary data are presented to quantify some performance factors concerned. We have not heard about inserting data method which purpose is such as index and content information insertion.

This paper presents a new general distance measure that not only can be used in a vector quantization (VQ) of line spectrum frequency (LSF) parameters but also performs well in a LSF transformed domain. The new distance is based on the spectral sensitivity of LSFs and their transformed coefficients. In addition, a fix scaling vector is used to decrease the sensitivity of spectral error at higher frequencies. Experimental results have shown that the proposed distance measure leads to as good as or better performance of VQ compared to other methods in the field of LSF coding. The use of this distance as the weighting function of the LSF transformed parameters is also suggested.

This paper describes a spatial spectral subtraction method by using the complementary beamforming microphone array to enhance noisy speech signals for speech recognition. The complementary beamforming is based on two types of beamformers designed to obtain complementary directivity patterns with respect to each other. In this paper, it is shown that the nonlinear subtraction processing with complementary beamforming can result in a kind of the spectral subtraction without the need for speech pause detection. In addition, the optimization algorithm for the directivity pattern is also described. To evaluate the effectiveness, speech enhancement experiments and speech recognition experiments are performed based on computer simulations under both stationary and nonstationary noise conditions. In comparison with the optimized conventional delay-and-sum (DS) array, it is shown that: (1) the proposed array improves the signal-to-noise ratio (SNR) of degraded speech by about 2 dB and performs more than 20% better in word recognition rates under the conditions that the white Gaussian noise with the input SNR of -5 or -10 dB is used, (2) the proposed array performs more than 5% better in word recognition rates under the nonstationary noise conditions. Also, it is shown that these improvements of the proposed array are same as or superior to those of the conventional spectral subtraction method cascaded with the DS array.

Classification of the new and used bills using the spectral patterns of raw time-series acoustic data (observation data) poses some difficulty. This is the fact that the observation data include not only a bill sound, but also some motor sound and noise by a transaction machine. We have already reported the method using adaptive digital filters (ADFs) to eliminate the motor sound and noise. In this paper, we propose an advanced technique to eliminate it by the neural networks (NNs). Only a bill sound is extracted from observation data using prediction ability of the NNs. Classification processing of the new and used bills is performed by using the spectral data obtained from the result of the ADFs and the NNs. Effectiveness of the proposed method using the NNs is illustrated in comparison with former results using ADFs.

This paper proposes an adaptive noise canceller with low signal-distortion in the presence of crosstalk. The proposed noise canceller has two pairs of cross-coupled adaptive filters, each of which consists of the main filter and a sub filter. The signal-to-noise ratios (SNRs) of the primary and the reference signals are estimated by the sub filters. To reduce signal distortion at the output of the adaptive noise canceller, the step sizes for coefficient adaptation in the main filters are controlled according to the estimated SNRs. Computer simulation results show that the proposed noise canceller reduces signal distortion in the output signal by up to 15 dB compared to the conventional noise canceller.

In this paper, a new gradient-based adaptive algorithm for the estimation of discrete Fourier coefficients (DFC) of a noisy sinusoidal signal is proposed based on a summed least mean squared error criterion. This algorithm requires exactly the same number of multiplications as the conventional LMS algorithm, and presents much improved performance in both white and colored noise environments at the expense of some additional memories and additions only. We first analyze the performance of the conventional LMS algorithm in colored additive noise, and point out when its performance deteriorates. Then, a summed least mean squared error criterion is proposed, which leads to the above-mentioned new gradient-based adaptive algorithm. The performance of the proposed algorithm is also analyzed for a single frequency case. Simulation results are provided to support the analytical findings and the superiority of the new algorithm.

This paper proposes adaptive line enhancers with new coefficient update algorithms on the basis of least-square-error criteria. Adaptive algorithms by least-squares are known to converge faster than stochastic-gradient ones. However they have high computational complexity due to matrix inversion. To avoid matrix inversion the proposed algorithms adapt only one coefficient to detect one sinusoid. Both FIR and IIR types of adaptive algorithm are presented, and the techniques to reduce the influence of additive noise is described in this paper. The proposed adaptive line enhancers have simple structures and show excellent convergence characteristics. While the convergence of gradient-based algorithms largely depend on their stepsize parameters, the proposed ones are free from them.

In this paper, we propose a new structure of blind equalizer and its cost function. The proposed cost function is a quadratic form and has the unique solution. In addition, the proposed scheme can employ iterative algorithms which achieve less computational complexity and can be easily realized in real time processing. In order to verify the effectiveness of the proposed schemes, several computer simulations including a 64-QAM signal equalization have been shown.

This paper presents an approach to the blind identification of multichannel communication systems by using partial knowledge of the channel. The received signal is first processed by a filter constructed by the known component of the channel and then a blind identification algorithm based on the second-order statistics is applied to the filtered signal. It is shown that, if the unknown component satisfies the identifiability condition, the channel can be identified even though the channel does not satisfy the identifiability condition. Simulation results are presented to show the performance of the proposed approach. A comparison to the existing approaches is also presented.

In some applications, such as the echo cancellation problem of satellite-linked communication channels, there occurs a problem of estimation of a long impulse response, which consists of a long flat delay and a short dispersive response region. In this paper, it is shown that the use of the adaptive algorithm based on the frequency domain sampling theorem enables efficient identification of the long impulse response. The use of the proposed technique can lead to the reduction of both the number of adaptive weights and the complexity of flat delay estimation.

This paper proposes and investigates the adaptive single-user receiver with co-channel interference (CCI) canceller based on orthogonalizing matched filter (OMF) using the multi-dimensional (multi-D) lattice filters for DS/CDMA in a multipath environment. A conventional single-user receiver using OMF cannot correctly cancel CCI in the presence of multipath in a channel, because the desired user's signal component and other users' intersymbol interference (ISI), due to multipath, still remain at the output of OMF, and then a correct replica of CCI cannot be generated. The proposed receiver can solve this problem because a multi-D IIR lattice filter can distinguish the desired user's signal component from the other users' ones and remove the ISI of all users' signals by Gram-Schmidt orthogonalization. And a multi-D FIR lattice filter can selectively regenerate undesired users' signal components which contain neither the desired user's signal component nor ISI. The proposed receiver based on multi-D lattice filters can apply orthogonality property of the OMF to not only a direct wave but also to delayed waves in a multipath channel. Since the proposed receiver does not have to know spreading codes of multiusers except the desired user, it will be applicable at a mobile terminal in the forward link in a multipath environment. Computer simulations show the proposed receiver have capability to cancel the CCI and converges fast in a multipath channel.

In 2001, the DEMETER micro-satellite will be launched to perform Detection of Electro-Magnetic Emissions Transmitted from Earthquake Regions. Its main scientific objective is related to the investigation of the ionospheric perturbations due to the seismic and volcanic activity. A system allowing an onboard identification and characterization of spatially and temporally coherent structures associated with the measurement of one or several electromagnetic wave field components is used. It is based on neural networks. The choice and training of the neural network are done on the ground from available waveforms. The parameters of the neural network system are then transmitted to the satellite. This reconfiguration process can be repeated whenever necessary during the space mission. Details about the functioning and coding optimization for DSP implementation is presented. The first function of this system which will be performed on the satellite DEMETER is the real-time identification and characterization of whistler phenomena. An application to the analysis of such phenomena observed in data from the AUREOL-3 satellite is exposed.

We proposed a new model for non-stationary time series analysis based on the IAR (inhomogeneous autoregressive) model, and a method for model parameter estimation when the set of basis is given. In this paper, we further propose a method for parameter estimation including that of basis set: we set a new condition that power of the input sequence is concentrated in low-frequency domain, and developed an iterative estimation method. We firstly select an initial set of basis, from which new sets are created in order to minimize the difference between the model and data. Among new sets of basis, we select a good one that gives minimum standard deviation of estimated frequencies.

An approach to an MPEG decoding system with reduced memory capacity will be presented. This method relies on the simple technique of one-dimensional DPCM to recompress reconstructed Macro Block (MB) prior to being stored on frame memory. Simulation results suggest that image quality is subjectively acceptable when using approximately one-half of the memory size required by that of conventional decoder. The degradation in the signal-to-noise ratio introduced by this compression method ranged from 0.1 dB to 0.7 dB for MPEG MP@ML standard test sequences at 4 Mbps. This technique can be implemented to achieve a cost effective MPEG decoder.

In the literature [9], the optimum discrete interpolation of one-dimensional signals is presented which minimizes various measures of approximation error simultaneously. In the discussion, the ratio λ of the weighted norm of the approximation error and that of the corresponding input signal plays an essential role to determine the structure of the set of signals. However, only the upper bound of λ is provided in [9]. In this paper, we will present more exact and systematic discussion of the optimum discrete interpolation of one-dimensional signals which minimizes various measures of approximation error at the same time. In this discussion, we will prove that the exact value of λ is identical with the upper limit, for ω (|ω| π), of the largest eigen value of a matrix including the weighting function W(ω) and the Fourier transforms of the optimum interpolation functions. Further, we will give a sufficient condition for W(ω) under which the ratio λ is equal to one, where the approximation error, if it is interpolated by sinc, is included in the set of band-limited signals defined by W(ω). Finally, as application of the presented approximation, we will propose a direction to interactive signal processing on Internet and a transmultiplexer system included in it. The transmultiplexer system included in this discussion can realize flexible arrangement of sub-bands which is inevitable in realizing the above proposal on interactive signal processing.

It is an important problem in signal processing, system realization and system identification to find linear discrete-time systems which are consistent with given covariance parameters. This problem is formulated as a problem of finding discrete-time positive real functions which interpolate given covariance parameters. Various investigations have yielded several significant solutions to the problem, while there remains an important open problem concerning the McMillan degree. In this paper, we use more general input-output characteristics than covariance parameters and consider finding discrete-time positive real matrix functions which interpolate such characteristics. The input-output characteristics are given by the coefficients of the Taylor series at some complex points in the open unit disk. Thus our problem is a generalization of the interpolation problem of covariance parameters. We reduce the problem to a directional interpolation problem with a constraint and develop the solution by a state-space based new approach. The main results consist of the necessary and sufficient condition for the existence of the discrete-time positive real matrix function which interpolates the given characteristics and has a limited McMillan degree, and a parameterization of all such functions. These are a contribution to the open problem and a generalization of the previous result.

We propose manifold piecewise constant systems (ab. MPC) and consider basic phenomena: the 2-D, 3-D and 4-D MPCs exhibit limit-cycle, line-expanding chaos and area-expanding chaos, respectively. The righthand side of the state equation is piecewise-constant, hence the system dynamics can be simplified into a piecewise-linear return map which can be expressed explicitly. In order to analyze the piecewise-linear return map, we introduce an evaluation function for the piecewise-linear return map and give theoretical evidence for chaos generation. Also the chaotic behaviors are demonstrated in the laboratory.

Fractal immittance, expressed by an admittance s^a (0<|a|<1), is simulated by the analog circuits composed of finite numbers of conventional elements, resistance R, capacitance C and inductance L, based on the distributed-relaxation-time models. The correlation between the number of R-C or R-L pairs and the optimum pole interval to give the widest bandwidth is estimated for each a-value by the numerical calculation for each circuit against a given criterion with respect to the phase angle. It is found that the bandwidth of 5 decades with a phase-angle error of 1 can be composed for |a|=0.1-0.9 using eighteen pairs or less of the elements.

In this paper we propose an algorithm for generating maximum weight independent sets in a circle graph, that is, for putting out all maximum weight independent sets one by one without duplication. The time complexity is O(n³ + β ), where n is the number of vertices, β output size, i. e. , the sum of the cardinalities of the output sets. It is shown that the same approach can be applied for spider graphs and for circular-arc overlap graphs.

Two classes of nonlinear feedforward logic (NLFFL) pseudonoise (PN) code generators based on the use of AND and majority logic (ML) gates are compared. Cross-correlation and code-division multiple-access (CDMA) properties of properly designed NLFFL sequences are found to be comparable with the properties of well-known linear PN codes. It is determined that code design employing ML gates with an odd number of inputs is easier compared with designing with AND gates. This is especially true when the degree of nonlinearity is large, since the nonbalance problem, e. g. , at the output of an AND gate, can be avoided. ML type sequences are less vulnerable to correlation attack and jamming by the m-sequence of an NLFFL generator

Petri net is a graphical and mathematical modeling tool for discrete event systems. This paper treats analysis problems of time Petri nets. In this model, a minimal and a maximal firing delays are assigned to each transition. If a transition is 'enabled' it can fire after minimal delay has passed and must fire before maximal delay has elapsed. Since time Petri net can simulate register machines, it has equivalent modeling power to that of Turing machine. It means, however, that most of the analysis problems of time Petri nets with general net structures are undecidable. In this paper, net structures are restricted to a subclass called partially ordered condition (POC) nets and dissynchronous choice (DC) nets. Firing delays are also restricted to satisfy 'static fair condition' which assures chance to fire for all transitions enabled simultaneously. First, a sufficient condition of liveness of time POC net with the static fair condition is derived. Then it is shown that liveness of time DC net with static fair condition is equivalent to liveness of the underlying nontime net. This means that liveness problem of this class is decidable. Lastly, liveness problem of extended free choice (EFC) net is shown to be decidable.

Lossless block transforms and filter banks that map integers to integers are important for unified lossless and lossy image coding systems. In this paper, we present simple yet effective methods for designing lossless versions of block transforms and FIR filter banks. First, an N-point lossless transform and a lossless interpolative prediction are introduced. Next, we demonstrate that filter banks can be decomposed into 2-point transforms or interpolative predictions. Lastly, lossless versions of block transforms and filter banks are obtained by replacing every constituent module by the corresponding lossless version. Lossless versions of 8-point discrete cosine transform (DCT), 8-point Walsh-Hadamard transform (WHT) and several filter banks are designed and their lossless compression performance is evaluated.

In this paper, we propose a new algorithm of enhancing covariance matrix estimate to be used for estimating the directions-of-arrival (DOAs) of multiple incoherent signals incident on a uniform circular array. The underlying covariance matrix possesses a special theoretical property such as having spatial stationarity. The proposed enhancement approach based on the use of this property is found to provide improved DOA estimates in comparison to the unenhanced MUSIC for narrowband incoherent signals.

This paper presents a novel dead-beat synchronization scheme and applies it to communications in discrete-time chaotic systems. A well-known Henon system is considered as an illustrative example. In addition, a Henon-based image processing application effectively exploits the proposed scheme's effectiveness.

In this letter, the novel mapping network named self-organizing relationship (SOR) network, which can approximate the desired I/O relationship by employing the modified Kohonen's learning law, is proposed. In the modified Kohonen's learning law, the weight vectors are updated to be attracted to or repulsed from the input vector.