Image segmentation is a critical step of image analysis. Segmentation evaluation is an effective procedure for studying the performance of segmentation techniques, in which quality measure plays an important role. This paper presents a group of new objective quality measures for egmentation evaluation and compares their performances. In addition, to verify the effectiveness of these new measures, an appropriate classification of segmentation is proposed. According to this classification, several representative algorithms from different categories are selected for comparison testing. Some valuable results are obtained and presented.

The characteristics of 2-D convolutions with big kernel in low-rank are analysed, and a fast algorithm is given. Then a systolic array implementation, which is derived by a three-stage dependence-graph-based mapping approach, is presented. It is shown that the architecture has a high efficiency for parallel processing and a nearly linear speed-up.

By comparison with constraint satisfaction networks, this paper presents an essential frame of the logical theory for continuous-state neural networks, and gives the quantitative analyzing method for contradiction. The analysis indicates that the basic reason for the alternation of the logical states of the neurons is the existence of superior contradiction inside the networks. The dynamic process for a neural network to find a solution corresponds to eliminating the superior contradiction.

The paper presents a total parallell neural network for real-tune computation of SVD. Its boundedness and stability of the dynamical system are studied. Computer simulation is also given. The results of the theoretical analysis and simulation illustate that this neural network is feasible and efficient for SVD.

The data assciation problem is one of the key problems of multitarget tracking in dense multiple return environments. By constructing a suitable energy function, the average values of a Boltzmann machine (T = 1) are approximately equal to the association probabilities. Then, a new method for computing association probabilities using mean-field approximation network is presented. The simulations show that this method is effective.

The paper proposes an estimation method of blocking in multistage interconnection networks serving multirate traffics. It can be used in multirate curcuit swithcing and ATM switching systems. In the paper, an analysis of blocking is given. The accuracy of the model is tested by simulation and some factors which influance the blocking are commpared. The study shows that the modular size and service bandwidth play an important role in the blocking behavior. Compared with non-blocking design, blocking design can reduce the complexity of harhware considerablly.

In ATM networks, bursty sources can be described as the Interrupted Bernoulli Pro-cess(IBP). With the use of the thin process theory, the Probability Generating Function(PGF) of the IBP is obtained. An iterative algorithm, which can be used to calculate the IBP probability distribution, is presented. The bursty source s equivalent description is discussed. It is proposed that the leaky bucket output process can be approximately described as IBP. The accuracy of the analytical results has been largely validated by means of the simulation approach. Moreover, how to improve its accuracy is discussed. The smoothing function of the leaky bucket algorithm is quantitatively analyzed.

This paper analyzes a queue model of the polling system with limited service (K = 1) in discrete time. By the imbedded Markov chain theory and the generating function method, the mean values of queue length and message waiting time are explicitly obtained. The results obtained by H. Tagai (1985) are revised.

This paper presents a real-time implementation of 4.2Kb/s CELP speech coding on single DSP. An algorithm reducing search complexity for adaptive codebook is suggested, the solving method, which changes a parameter into LSP parameter, is discussed. The real-tune implementation procedure of this coding on a commercial development board with a single TMS320C30 is discribed.

This paper presents lower bounds and upper bounds for the mean value and variance of distance distribution of non-linear codes in GF(q). By presenting several examples, it is shown that these bounds could be achieved.

In order to resist the anti-radar missiles, this paper presents a method for designing a group-complementary phase-coded signal set in which each signal has a characteristic of low probability intercept and is orthogonal to other signal. Meanwhile, on the basis of the orthognality of this kind of group-complementary phase-coded signals, this paper also presents a new method for Doppler phase compensation. Finally, this paper performs computer simulations to confirm the presented methods.

This paper presents a novel technique for directly IF sampling and digitally obtaining the inphase (I) and quadrature (Q) components of the IF signal. Because all subsequent processing and detection are accomplished in the digital domain, it can provide even more unproved I and Q signals and improve the overall performance greatly. Finally an experimental scheme is provided.

The characteristics method is introduced to analyze the transient sensitivity of transimission lines in high-speed VLSI circuits with respect to the transmission line parameters or terminal parameters, supplying a useful tool for the optimization design of interconnects in VLSI circuits. The characteristics method differs from the existing sensitivity analysis method based on the numerical inverse Laplace Transform(NILT) technique mainly in that the characteristics method can deal with nonlinear terminals and arbitrary transmitted signals.

Factor analysis method is firstly used to establish 5 parameter statistical model of microwave FETs. This statistical model is compared with the one established by principal component analysis method. The results show that the statistical model by factor analysis is more accurate than that by principal component analysis

This paper presents the analysis of open microstrip structures by using diakoptic method of lines combined with periodic boundary conditions (PBC). The parameters of microstrip patch are obtained from patch current excited by plane wave. Impedance matrix elements are computed by using fast Fourier transform (FFT), and reduced equation is solved by using diakoptic technique. Consequently, the computing time is reduced significantly. The convergence property of simulating open structure by using PBC and the comparison of the computer time between using PBC and usual absorbing boundary condition(ABC) show the validity of the method proposed in this paper. Finally, the resonant frequency of a microstrip patch is computed. The numerical results obtained are in good agreement with those published.

Based on the method of separating variables and Muller s calculating roots, the propagation characteristics of hybrid modes in homogeneous and inhomogeneous circular Faraday chi-rowaveguides are examined in detail. Numerical results are presented to demonstrate the effects of different constitutional parameters on the mode bifurcation, symmetry and nonreciprocity, backward-wave and attenuation of hybrid modes, which are different from the cases of ordinary gyrotropic and reciprocal chiral waveguides.

The virtual ray method for treating HF electromagnetic scattering problems is derived from the plane wave of free space, by use of which the plane wave scattering by an impedance wedge is studied. In the treatment the concept of generalized circle is introduced so that the complete amplitude function is obtained which is a component of the solution. And a reasonable physical interpretation of the term w 2, which was nelected previously, is given. The calculated result agrees well with the analytical soluton obtained by G. D. Maliuzhinets(1958).

Direct TM-Scattering problem of two-dimensional inhomogeneities buried in a loosy stratified medium is presented. Analysis work mainly consists of two parts: one is to derivate the Green s function of a filament buried in a loosy stratified medium, another is to constitute a electric field integral equation of an equivalent current caused by the differences between the inhomogeneities and the stratified medium. Based on these works, illustrative numerical results are given to model inhomogeneous underground tubes embedded in a stratified medium, and to describe the scattering field affected by different factors such as permittivity distribution, dimension and buried depth of the inhomogeneities and so on.

This paper addresses the effect of nonlinearity of RCAM on its recall performance with the help of signal-to-noise analysis and presents a new nonlinear function. Theoretical analysis and computer simulations show that the presented memory model offers more satisfactory performance.

An image processing system under MS-Windows environment has been developed on micro-computer in order to facilitate the data analyzing of all-sky auroral image recorded in polar region. The system has realized: information extraction of aurora s shape, position, motion and brightness on geomagnetic coordinate.

The principles, properties of Hilbert transform and wavelet transform and their applications to extracting signal envelope are discussed respectively. Some comparisons are made to illustrate that modulated Gaussian wavelet transform not only can orthogonalize signal, but also has band-pass filtering function, so it is a more ideal method for extracting signal envelope in contrast with Hilbert transform.

This paper presents an Azimuth Multi-look Processor (AMP). It is a major part of airborne Synthetic Aperture Radar (SAR) real-time digital image processor. And its concurrent and real-time processing abilities are discussed. A feasible configuration of the AMP is given. Finally, the characteristics and applications of the multitask multiprocessor AMP are also presented.

This paper presents an idea for the average brightness temperature of antenna aperture and its expression for a lossless antenna with matched and unmatched noise sources. The expressions show that the average brightness temperature of antenna aperture relates with three factors: coefficient of useful area for antenna, noise temperature of noise sources and reflective coefficient of noise sources.

By analyzing the scattered fields from zone plate reflectors, the focal field distributions of Presnel zone plate antenna are investigated in the view of receiving antenna analysis.

A high accuracy stable Dispersive Boundary Condition(DBC) is presented, which can be used to model wave propagation in transmission lines by the finite-difference time-domain method. The differential factor in the boundary condition is replaced by a new two-order difference scheme. The new DBC has the same absolute stability as that proposed by P. Y. Zhao, et al. (1994), but it has a much better absorbing performance.

A new method for analysing SOI (Silicon on Insulator) Sructure Mach-Zehnder interferometric modulator by using finite element method is put forward. On the basis of the theory of the single-mode SOI rib optical waveguides with large cross-section, the electro-optic modulating mechanism of the modulator is investigated by using the plasma dispersion effect, and the electrical characteristics of the device is analysed by using the finie element method at p⁺n junction large injected. So the method provides a basis of the theory for the device to be fabricated.