Mohammed BENNAMOUN Boualem BOASHASH
We previously proposed a robust hybrid edge detector which relaxes the trade off between robustess against noise and accurate localization of the edges. This hybrid detector separates the tasks of localization and noise suppresion between two sub-detectors. In this paper, we present an extension to this hybrid detector to determine its optimal parameters, independently of the scene. This extension defines a probabilistic cost function using for criteria the probability of missing an edge buried in noise and the probability of detecting false edges. The optimization of this cost function allows the automatic selection of the parameters of the hybrid edge detector given the height of the minimum edge to be detected and the variance of the noise, σ2n. The results were applied to the 2D case and the performance of the adaptive hybrid detector was compared to other detectors.
Mohammed BENNAMOUN Boualem BOASHASH
Within the framework of a previously proposed vision system, a new part-segmentation algorithm, that breaks an object defined by its contour into its constituent parts, is presented. The contour is assumed to be obtained using an edge detector. This decomposition is achieved in two stages. The first stage is a preprocessing step which consists of extracting the convex dominant points (CDPs) of the contour. For this aim, we present a new technique which relaxes the compromise that exists in most classical methods for the selection of the width of the Gaussian filter. In the subsequent stage, the extracted CDPs are used to break the object into convex parts. This is performed as follows: among all the points of the contour only the CDPs are moved along their normals nutil they touch another moving CDP or a point on the contour. The results show that this part-segmentation algorithm is invariant to transformations such as rotation, scaling and shift in position of the object, which is very important for object recognition. The algorithm has been tested on many object contours, with and without noise and the advantages of the algorithm are listed in this paper. Our results are visually similar to a human intuitive decomposition of objects into their parts.
Mitsuhiko YAGYU Akinori NISHIHARA Nobuo FUJII
This paper presents a method to analyze and minimize output errors of 2-D non-separable FIR filters with finite wordlength. Finiteness in the wordlength causes output errors, which can be analyzed in the frequency domain when the statistics of input signals are known. The output errors can be minimized by optimizing responses corresponding to all levels of input impulses. A new ROM-based filter structure is proposed in which the optimized impulse responses are stored in the ROM. The output signals are generated by superposing the impulse responses corresponding to the input levels. Many simulation results confirm that the output signals of the proposed filters have far less errors compared to conventional filters. The hardware size of the ROM-based filters is estimated and compared with that of conventional structures. The proposed structures are more effective than the conventional ones especially when the signal wordlength is short.
Shuitsu MATSUMURA Fumihiko MURATA Tsuyoshi TAKEBE
This paper describes a design technique of perfect reconstruction (PR) two-channel IIR filter bank. M.J.T. Smith et al., gave two types of PR IIR filter bank systems. One is the system such that the analysis and synthesis filters with nonlinear phase are implemented with all-pass polyphase filters and satisfy the power complementary condition approximately. The other is the system such that all the analysis and synthesis filters have liner phase responses and do not satisfy the power complementary condition. To improve coding performance, we propose a filter bank system such that all the analysis and synthesis filters have linear phase and satisfy the power complementary condition approximately.
Xi ZHANG Toshinori YOSHIKAWA Hiroshi IWAKURA
This paper presents a new method for constructing orthonormal wavelet bases with vanishing moments based on general IIR filters. It is well-known that orthonormal wavelet bases can be generated by paraunitary filter banks. Then, synthesis of orthonormal wavelet bases can be reduced to design of paraunitary filter banks. From the orthonormality and regularity of wavelets, we derive some constraints to IIR filter banks, and investigate relations between the constrained filter coefficients and its zeros and poles. According to these relations, we can apply Remez exchange algorithm in stopband directly, and formulate the design problem in the form of an eigenvalue problem. Therefore, a set of filter coefficients can be easily computed by solving the eigenvalue problem, and the optimal filter coefficients with an equiripple response can be obtained after applying an iteration procedure. The proposed procedure is computationally efficient, and the number of vanishing moments can be arbitrarily specified.
Hideyuki IMAI Akira TANAKA Masaaki MIYAKOSHI
A lot of optimum filters have been proposed for an image restoration problem. Parametric filter, such as Parametric Wiener Filter, Parametric Projection Filter, or Parametric Partial Projection Filter, is often used because it requires to calculate a generalized inverse of one operator. These optimum filters are formed by a degradation operator, a covariance operator of noise, and one of original images. In practice, these operators are estimated based on empirical knowledge. Unfortunately, it happens that such operators differ from the true ones. In this paper, we show the unified formulae of inducing them to clarify their common properties. Moreover, we investigate their properties for perturbation of a degradation operator, a covariance operator of noise, and one of original images. Some numerical examples follow to confirm that our description is valid.
Xiaoxing ZHANG Masahiro IWAHASHI Noriyoshi KAMBAYASHI
In this paper a novel narrow-band bandpass filter with an output pair of analytic signals is presented. Since it is based on the complex analog filter, both synthesis and response characteristics of this filter are different from conventional bandpass filters. In the design of this filter, the frequency shift method is employed and the conventional lowpass to bandpass frequency transformation is not required. The analysis and examples show that the output signal pair of the proposed filter possesses same filtering characteristics and a 90 degree phase shifting characteristics in the passband. Therefore, the proposed filter will be used for a single sideband (SSB) signal generator without quadrature generator.
A new method is proposed for estimating a single complex sinusoid and its parameters (frequency and amplitude) from measurements corrupted by white noise. This method is called the ECKF-SVD method, which is derived by applying an extended complex Kalman filter (ECKF) to a nonlinear stochastic system whose state variables consist of the AR coefficient (a function of frequency) and a sample of the original signal. Proof of the stability is given in the case of a single sinusoid. Simulations demonstrate that the proposed ECKF-SVD method is effective for estimating a single complex sinusoid and its frequency under a low signal-to-noise ratio (SNR). In addition, the amplitude estimation by means of the ECKF-SVD method is also discussed.
Allan KARDEC BARROS Noboru OHNISHI
Event-related are the kind of signals that are time-related to a given event. In this work, we will study the effect of bias and overlapping noise on Fourier linear combiner (FLC)-based filters, and its implication on filtering event-related signals. We found that the bias alters the weights behaviour, and therefore the filter output, and we discuss solutions to the problem of spectral overlap.
This paper compares signal classification performance of multilayer neural networks (MLNNs) and linear filters (LFs). The MLNNs are useful for arbitrary waveform signal classification. On the other hand, LFS are useful for the signals, which are specified with frequency components. In this paper, both methods are compared based on frequency selective performance. The signals to be classified contain several frequency components. Furthermore, effects of the number of the signal samples are investigated. In this case, the frequency information may be lost to some extent. This makes the classification problems difficult. From practical viewpoint, computational complexity is also limited to the same level in both methods.IIR and FIR filters are compared. FIR filters with a direct form can save computations, which is independent of the filter order. IIR filters, on the other hand, cannot provide good signal classification deu to their phase distortion, and require a large amount of computations due to their recursive structure. When the number of the input samples is strictly limited, the signal vectors are widely distributed in the multi-dimensional signal space. In this case, signal classification by the LF method cannot provide a good performance. Because, they are designed to extract the frequency components. On the other hand, the MLNN method can form class regions in the signal vector space with high degree of freedom.
Tohru IKEGUCHI Kazuyuki AIHARA
In this paper, we propose a new strategy of estimating correlation dimensions in combination with the method of surrogate data, which is a kind of statistical control usually introduced to avoid spurious estimates of nonlinear statistics, such as fractal dimensions, Lyapunov exponents and so on. In the case of analyzing time series with the method of surrogate data, it is desirable to decide values of estimated nonlinear statistics of the original data and surrogate data sets as exactly as possible. However, when dimensional analysis is applied to possible attractors reconstructed from real time series, it is very dangerous to decide a single value as the estimated dimensions and desirable to analyze its scaling property for avoiding spurious estimates. In order to solve this defficulty, a dimension estimator algorithm and the method of surrogate data are combined by introducing Monte Carlo hypothesis testing. In order to show effectiveness of the new strategy, firstly artificial time series are analyzed, such as the Henon map with additive noise, filtered random numbers and filtered random numbers transformed by a static monotonic nonlinearity, and then experimental time series are also examined, such as wolfer's sunspot numbers and the fluctuations in a farinfrared laser data.
Hironori TOKUNO Ole KIRKEBY Philip A. NELSON Hareo HAMADA
We present a very fast method for calculating an inverse filter for audio reproduction system. The proposed method of FFT-based inverse filter design, which combines the well-known principles of least squares optimization and regularization, can be used for inverting systems comprising any number of inputs and outputs. The method was developed for the purpose of designing digital filters for multi-channel sound reproduction. It is typically several hundred times faster than a conventional steepest descent algorithm implemented in the time domain. A matrix of causal inverse FIR (finite impulse response) filters is calculated by optimizing the performance of the filters at a large number of discrete frequencies. Consequently, this deconvolution method is useful only when it is feasible in practice to use relatively long inverse filters. The circular convolution effect in the time domain is controlled by zeroth-order regularization of the inversion problem. It is necessary to set the regularization parameter β to an appropriate value, but the exact value of β is usually not critical. For single-channel systems, a reliable numerical method for determining β without the need for subjective assessment is given. The deconvolution method is based on the analysis of a matrix of exact least squares inverse filters. The positions of the poles of those filters are shown to be particularly important.
Hirofumi NAKAJIMA Masato MIYOSHI Mikio TOHYAMA
The Multiple input-output INverse/filtering Theorem (MINT) proves that N + 1 inverse filters are necessary to precisely control sound at N points in a space, and gives the minimum orders of such filters. In this paper, we propose the Indefinite MINT Filters (IMFs) for adding one or more control points to the above framework without increasing the number of inverse filters. Although the controllability of the new point is not sufficient, that of the other points is still maintained high enough by the principle of the MINT. In a two point sound control (using two inverse filters), the IMFs could reduce the squared error to the desired sound up to - 10 dB at the second point which is not controlled by the MINT.
Yasuo KOKUBUN Shigeru YONEDA Hiroaki TANAKA
The temperature dependence of the central wavelength of narrow-band filters is a serious problem for the dense WDM systems. In this study, we realized a temperature independent narrow-band filter at 1.3 µm wavelength. First, we designed an athermal waveguide in which optical path length is independent of temperature by using a finite element method. Using this athermal waveguide, we designed and fabricated a ring resonator. As a result, we successfully decreased the temperature coefficient of central wavelength to 710-4 nm/K, which is 7% of conventional SiO2 waveguide filters and 0.7% of conventional semiconductor waveguide filters.
Atsushi YAMAGUCHI Hiroyuki FURUYA Kensaku FUJII Juro OHGA
The filtered-x algorithm, which is widely applied to active noise control system, requires setting a small step gain. Such a small step gain reduces the noise reduction effect when the alogrithm is implemented by fixed point processing. This paper presents an experimental result that the 'polarized-g' individually normalized least mean square (INLMS) algorithm can provide almost the same noise reduction effect even in the fixed point processing of 16 bits as that in floating point processing.
A new method is proposed for recovering an unknown source signal ,which is observed through two unknown channels characterized by non-minimum phase FIR filters. Conventional methods cannot estimate the non-minimum phase parts and recover the source signal. Our method is based on computing the eigenvector corresponding to the smallest eigenvalue of the input correlation matrix and using the criterion with the multi-channnel inverse filtering theory. The impulse responses are estimated by computing the eigenvector for all modeling orders. The optimum order is searched for using the criterion and the most appropriate impulse responses are estimated. Multi-channel inverse filtering with the estimated impulse responses is used to recover the unknown source signal. Computer simulation shows that our method can estimate nonminimum phase impulse responses from two reverberant signals and recover the source signal.
Saed SAMADI Akinori NISHIHARA Nobuo FUJII
It is shown that two-dimensional linear phase FIR digital filters with various shapes of frequency response can be designed and realized as modular array structures free of multiplier coefficients. The design can be performed by judicious selection of two low order linear phase transfer functions to be used at each module as kernel filters. Regular interconnection of the modules in L rows and K columns conditioned with boundary coefficients 1, 0 and 1/2 results in higher order digital filters. The kernels should be chosen appropriately to, first, generate the desired shape of frequency response characteristic and, second, lend themselves to multiplierless realization. When these two requirements are satisfied, the frequency response can be refined to possess narrower transition bands by adding additional rows and columns. General properties of the frequency response of the array are investigated resulting in Theorems that serve as valuable tools towards appropriate selection of the kernels. Several design examples are given. The array structures enjoy several favorable features. Specifically, regularity and lack of multiplier coefficients makes it suitable for high-speed systolic VLSI implementation. Computational complexity of the structure is also studied.
Kawori TAKAKUBO Hajime TAKAKUBO Shigetaka TAKAGI Nobuo FUJII
Analog inverter is one of the most useful building blocks in analog circuits. This paper proposes an analog inverter consisting of a p-channel MOS (PMOS) and an n-channel MOS (NMOS) inverter and presents an application to all-pass filter realizations. The proposed circuit has a wide dynamic range by combining PMOS and NMOS inverters. When the proposed analog inverter is applied to an all-pass filter, the circuit configuration becomes simpler and occupies less chip area and power consumption.
The numerical properties of the recursive least squares (RLS) algorithm and its fast versions have been extensively studied. However, very few investigations are reported concerning the numerical behavior of the predictor based least squares (PLS) algorithms that provide the same least squares solutions as the RLS algorithm. This paper presents a comparative study on the numerical performances of the RLS and the backward PLS (BPLS) algorithms. Theoretical analysis of three main instability sources reported in the literature, including the overrange of the conversion factor, the loss of symmetry and the loss of positive definiteness of the inverse correlation matrix, has been done under a finite-precision arithmetic. Simulation results have confirmed the validity of our analysis. The results show that three main instability sources encountered in the RLS algorithm do not exist in the BPLS algorithm. Consequently, the BPLS algorithm provides a much more stable and robust numerical performance compared with the RLS algorithm.
Satoshi OKUDE Tetsuya SAKAI Masaaki SUDOH Akira WADA Ryozo YAMAUCHI
A novel technique is proposed to fabricate a chirped fiber Bragg grating utilizing thermal diffusion of core dopant. The chirped grating is written with a uniform period by using UV exposure technique in the fiber whose effective index of the guided mode varies along its length. Thermal diffusion of the core dopant it employed to realize this change of the effective index. Through the thermal diffusion process, the effective index of the fiber decreases from its initial value. When the grating is written in the diffused core region, its reflection wavelength becomes shorter than that in the non-diffused region. The continuous change of effective index is required for making a chirped grating. The fiber is heated by a non-uniform heat source. When the uniform grating is written in this region, the reflection wavelength smoothly changes along the fiber length although the grating period is constant. By optimizing the fiber parameters to realize a highly chirped grating, we have obtained a typical one whose bandwidth is 14.1 nm at half maximum and maximum rejection in transmission is 29 dB. Additionally, the proposed method has an advantage to control the chirp profile with high mechanical reliability.