Wavelet Analysis and Its Applications, and Active Media Technology

Preface.

Conference Organization.

We specialize to some simple cases the algorithm for singularity detection in images from eigenvalues of the dual local autocovariance matrix. The eigenvalue difference, or "edginess" at a point, then reduces to a simple nonlinear function. We discuss the derivation of these functions, which provide low-complexity nonlinear edge filters with parameters for customization, and obtain formulas using the two smallest masks. We then exhibit their effects on six sample images.

The main features of a novel approach for audio signal encoding are described. The approach combines non-linear transform coding and structured approximation techniques, together with hybrid modeling of the signal class under consideration. Essentially, several different components of the signal are estimated and transform coded using an appropriately chosen orthonormal basis. Different models and estimation procedures are discussed, and numerical results are provided.

A new method of system identification based on wavelet analysis is proposed. An application of the formula to system identification of a health monitoring system is given.

The concept of Quaternion Wavelets is presented derived from the one of Clifford Wavelets as well as its possible use representing vortices in fluid mechanics, to do that we first show how vortices may be represented using quaternions.

This paper develops a wavelet method to fast estimate the eigenfunctions of Karhunen-Loeve expansion which is widely used to extract feature and simplify the model in signal processing, image analysis and dynamic system and so on. The integral equation is discretized by projecting the kernel function onto the wavelet space. The orthonormal wavelet bases are used in the paper. And the integral problem is transformed to the eigenvalue decomposition problem. If the kernel function satisfied the smooth requirement, the size of the matrix can be further reduced without sacrificing much accuracy. The numerical results are given to show the validity and effectiveness of the method.

Gabor function and log Gabor function were compared in this paper. The log Gabor wavelet was constructed and its performance was analyzed in the viewpoint of human visual system characteristics. Compared with Gabor wavelet, the excellent performance of log Gabor wavelet was shown. Finally, log Gabor wavelet was applied to image edge detection, and it is shown that using log Gabor wavelet can not only improve the detection properties, but also save about half the computational load compared with Gabor wavelet, in the similar detection effect case.

Edge detection is one of the most important starting steps in image processing. Many edge detection operators have been proposed based on regular convolution kernel, such as Sobel operator and Prewitt operator etc. This work summarizes the elementary edge types of binary image, based on which a novel approach for edge detection is proposed. As the simplest operator based on regular convolution kernel, its computational complexity is about one-fifth of Sobel operator's, it is well suitable for real-time applications. In respect of binary images, it can extract the edges accurately in noise-free case. In terms of accuracy and computational efficiency, the simplest operator performance is shown to be superior to those conventional edge detectors.

Traditional benchmarking methods for information retrieval (IR) are based on experimental performance evaluation^[1~14]. Although the metrics precision and recall can measure the effectiveness of a system, it cannot assess the underlying model. Recently, a theory of "aboutness" has been used for functional benchmarking of IR. Latest research shows the functionality of an IR model is largely determined by its retrieval mechanism, i.e., the matching function. In particular, containment and overlapping (either with or without a threshold value) are the core IR matching functions .The objective of this paper is to model the containment and overlapping matching functions using an aboutness-based framework and analyze them from an abstract and theoretical viewpoint. Separate aboutness relations for containment, pure-overlapping (i.e., without threshold) and threshold-overlapping matching functions are defined, and the sets of properties supported by them are derived and analyzed respectively. These three relations can be used to determine and explain the functionality of an IR system and how the functionality affects the performance of the system. Moreover, they can provide the design guidelines for new IR systems.

The construction method of a pair of biorthogonal interpolatory multiscaling functions is given. What is more, the corresponding biorthogonal interpolatory multiwavelets can also be constructed. Thereby, multiwavelets sampling theorem is established easily.

In this paper, 3-d wavelet moments are developed as new shape descriptors for 3-d objects. Compared with existed moments, they are not only robust to translation, rotation and scale, but also add the multi-resolution feature in the radical direction, which can handle noise to some extent and provide multi-level features to satisfy various requirements. At the same time, the computation of wavelet moments can be accelerated by using the fast algorithm of the spherical harmonics together with the Mallat algorithm of wavelets. Moreover, the orthonormality and the spatial completeness of the wavelet function make the reconstruction from the wavelet moments into reality.

Up to now, all the wavelet transforms are inner product representation of Hilbert space. In [1,2,3,4], based on the principle of mathematics and dialectical method, a novel method—VPWT (Vector Product Wavelet Transform) is proposed and studied, it is useful for wavelet theory and signal processing. In [5,6], vector wavelet functions, vector scaling functions and biased wavelet functions are put forward, they are suitable for signal processing. In this paper, combined [1,2,3,4] with [5,6], a new signal processing way is given out. Numerical tests of the new way by optimal control theory suggest that the new method considerably improve the representation capabilities of wavelet expansions.

In this paper, we have proved that a multivariable compactly supported polynomial spline on rectangular grids in ℝ^S as a scaling function of s dimensional wavelets must be in form of tensor product of univariable B-splines.

A orthogonal M-channel FIR multifilter banks can be expressed by its polyphase matrix, so the factorization of this system is reduced to the one of its corresponding polyphase matrix. Based on the singular value decomposition of the coefficient matrices of the polyphase matrix satisfied some conditions, we propose an order-one factorization form of this matrix. Furthermore, the factorization of the corresponding multifiter banks can be derived.

Riemenschneider and Shen constructed univariate interpolatory wavelet packets which provided a finer decomposition of the interpolation operator and a more detailed analysis of the structure of the given set of date. In this note, we extend this concept to the higher dimensions setting for arbitrary dilation matrices. This leads to a general way for the construction of multivariate interpolatory wavelet packets. Further, we also present the decomposition of multivariate interpolation wavelet packets.

In order to enhance speed of wavelet transform in signal processing, the author elaborates accelerative computing theory, algorithm for generalized fast wavelet transform and the fast lifted wavelet transform from the perspective of the multi-resolution analysis theory. The capability of accelerative algorithm is proved in theory. Apply this idea to multiresolution representation for medical image, the quality of image is retained and the time spend is saved effectively by this algorithm.

The classical Shannon sampling theorem has resulted in many applications and generalizations. From a multi-resolution point of view, it provides the Sinc function. In this paper, a general sampling theorem with compactly supported sampling function for the wavelet subspace is obtained. A method for efficient computation of inner products in multi-resolution analysis is proposed by using the Mallat algorithm.

This paper presents CL multiwavelet prefilter based on genetic algorithms. In this method, the fitness functions were selected for the different purposes of image processing. Image decomposition and reconstruction were used to construct the prefilters of images for CL multiwavelet transform (MWT) under the proper constraints. If image compression is needed, the fitness function is opted to obtain the prefilter adaptively via genetic algorithm for different images. It is shown that the image compression ratio can be improved if the prefilter is used. And if the purpose is to magnify images, the prefilter can be gained in the same way. Moreover, the quality of the image after magnification is superior to other methods.

According to the theory of 2D Lagrangian interpolation, interpolating wavelets be constructed. The wavelets have symmetry property; the scaling function and physical space representation are identical; expansion coefficients are easily computed.

In this paper, a novel WNN, multi-input and multi-output feedforward wavelet neural network is constructed. In the hidden layer, wavelet basis functions are used as activate function instead of the sigmoid function of feedforward network. The training formulas based on BP algorithm are mathematically derived and training algorithm is presented. This wavelet neural network could be used in multi-variable functional approximation, classification and pattern recognition.

A fast filter algorithm for real-time signal through wavelet transform is presented based on fast wavelet transform for 1D real-time signal. It is proved that the time complexity of this filter algorithm is irrelevant to the signal scale, but determined by the length of filter and the layers of wavelet decomposing. With the fast signal filter algorithm, a pipelining of data sampling and pre-processing is proposed. The fast filter algorithm is carried out to pre-process signal during the sampling interval so that there is no additional time spent. The pipelining has the same effect as conventional idea but economize the time spent on pre-processing to a great extent. This work lays a foundation for reducing time spent on signal processing and pattern recognizing.

We propose a new algorithm based on the optimality of orthogonal experimental design method and the abilities of memory in particles, which is called Orthogonal Particle Swarm Optimization (OPSO). Its characteristic is that initial particles of particle swarm are generated by orthogonal experimental design, so that these particles can be scattered uniformly over the feasible solution space; the particle swarm of the next generation is generated by means of memory. So the search space could be explored and exploited efficaciously. The OPSO is tested and compared with the standard PSO on four benchmark functions. The experimental results illustrate that the OPSO has the potential to achieve faster convergence and to find a better solution. In the end, the performance of the PSO is discussed with different settings of initial velocity.

This paper presents a novel motion estimation algorithm based on the continuous wavelet transform (CWT) and extended Kalman filter (EKF) models. The main goal is to detect and track moving objects in a scene, construct the trajectories. The algorithm uses a continuous wavelet transform (CWT) to analyze each frame. For motion estimation, we use an extended Kalman filter (EKF) as an estimator. It turns out that the CWT can be used efficiently in an EKF framework to perform detection and tracking. To provide the detection and tracking accuracy and efficiency, using multiscale edges detection (MED) firstly. Then, the CWT is used to map the input signal space to a motion parameters space. The estimates are combined over all orientations and scales using a coarse-to-fine refinement strategy to produce the accurate motion estimation.

This paper deals with the design and implementation of an image coding algorithm based on the integer wavelet transform(IWT) by the lifting scheme. IWT is more efficient than conventional float-point wavelet transform, its operation is faster. In case of hardware implementation, integer based arithmetic is much cheaper than those capable of handling floating points. In terms of memory usage, integer decomposition based on lifting scheme reduce the demands on the system by at least a factor two. For these reasons, the lossy image compression is considered. The image is decomposed by the IWT based on lifting scheme, coupled with modified EZW and adaptive range coder. Compared to Shaprio's EZW coder, the execution time of our algorithm is more than 1.5 times faster without degrading the performance.

A robust watermarking algorithm based on chaotic maps over the spatial domain is presented in this paper. Firstly, the original binary watermark is masked by a binary random sequence extracted from a chaotic trajectory and the host image is permuted through applying the Baker map. Secondly, the masked watermark is used to modify the 3rd 4th, or 5th bit of the corresponding pixels in the permuted host image. Finally, the permuted and watermarked image is inverse-permuted and the result is obtained. The simulation results have demonstrated that the scheme has a good transparence and resistant to be cropped. The extracted watermark from watermarked images are can be used to clearly identify the owner and still recognizable even the cropping ratio reaches 75%.

The scan imagery of electrical engineering drawings through preprocessing (binaryzation processing) and refinement processing, has been converted into a line pattern constituted by single pixels. Its compression algorithm is discussed in this paper in terms of single pixel graphs.

A new wavelet shrinkage de-noising algorithm based on the point singularity is presented. The approach determines the wavelet shrinkage threshold according to the singularity of image. The experimental results show that this new threshold achieves higher quality than Donoho's.

A color digital image watermarking algorithm, which is based on the zerotree structure of wavelet, is presented. The thresholds of significant coefficients are determined according to the capacity of watermarking and the characteristics of carrier images. The quantization step is dynamically adjusted. The watermarking is embedded into the Y luminance component of the original image's YIQ color space in DWT (Digital Wavelet Transform) domain. The embedded watermarking has double security. The algorithm is easy to implement with high robustness, and the watermark can be extracted without the original image.

Power quality monitors in being adopt the series-processing technique based on sampling-analyzing-sampling, skips over a lot of data between two-samplings. The paper puts forward method based on period field, process the data not only a major parameters of state standard, but also short-time and long-time issues, so it can realize real-time monitoring and quantificational analysis. Uses synchronized sampling data and error item to express non-synchronized sampling serial, then advances a iterative formula. Non-synchronized data is close to synchronized data by iterative operation. From the point of view of data of period signal, the method decreases and disappears even data leaking. In addition, the method decrease the amount of sampling points in one period, and sampling data of several sampling interval.

This paper is about seismic data compression and denoising by balanced orthogonal multiwavelet packet transform. Seismic data is compressed by selecting Lebrun's two order balanced orthogonal multiwavelet and the best base in the orthogonal multiwavelet packet, furthermore restored section is wavelet filtered section. Therefore, seismic data can be compressed effectively and signal-noise rate can be improved greatly at the same time. The result of the test shows that the proposed method has a few advantages over scalar wavelet method in terms of noise elimination and data compression.

The paper presents a new real-time modeling algorithm to simulate water wave. Two water wave stimulating functions, which take into account of water velocity, riverbed condition, and angle between two adjacent faces of riverbed, are given to serve as the stimulating source of water wave height field. Then a high efficient wave spread model, which is to stimulate the spread and reflection of wave, is proposed. The combination of stimulating function and wave spread model can stimulate the change of wave height of both stagnant water and flowing water. It is distinct from traditional methods that either construct special functions to model the shape of water wave or solve the complex Navier-Strokes equations. Texture mapping and Texture disturbing technology is used to simulate the water surface reflection effect. It turns out that our modeling algorithm satisfies the requirements of real time simulation of realistic water wave on generic PC platform, and comes up with new train of thought to stimulate water effect in complex environment.

In this paper, we propose a bivariate non-tensor product adaptive pre-wavelet neural network (APWNN), whose hidden layer consists of scale function nodes and three groups of wavelet function nodes. And we give a initialization and training algorithm for APWNN. At last a numerical example is given to validate the proposed APWNN.

New approach are presented for the design of pairs of wavelets with fractional filters where the two wavelets form an approximate Hilbert transform pairs. Wavelets with an arbitrary number of vanishing moments can be designed by using the new approach. Since polynomial filters is the especial setting of fractional filters, this design method can also be used to construct an approximate Hilbert transform pairs of wavelets bases with polynomial filters.

A novel approach of constructing wavelet packets with an arbitrary dilation matrix A is presented. The difference between the classical wavelet packets, instructed by Coifman et al., and ours is that we split the wavelet subspaces directly instead of using the low-pass and the high-pass filters associated with the multiresolution analysis. Furthermore, this method overcomes the difficulty of constructing non-orthogonal wavelet packets of the dilation matrix not equal to 21. Finally, we show how to construct various Riesz basis of L²(R^s) from the wavelet packets.

A new radar emitter recognition algorithm based on fuzzy comprehensive evaluation is presented by using the synthetic analysis method in modern mathematics. The detailed steps of the new recognition algorithm are also put forward. Moreover, the algorithm is analyzed and compared with classical statistical pattern recognition. The simulation results show that the performance of fuzzy comprehensive evaluation algorithm is much better than that of the classical method. The new approach is approved to be effective and feasible to radar emitter recognition.

In Embedded Zero-trees Wavelet Compression, there exist two problems, i.e. multi-scanning of wavelet coefficients and redundancy in the binary code. Considering it, this paper presents Improved EZW Image Coding Algorithm Based on zero-tree Index. The algorithm builds a new strategy with a from the bottom up scanning path by introducing the concept of zero-tree index, which reduces the times of repeated scanning and simplifies the computational complexity. We facilitate successive exclude scanning to get rid of the position information redundancy. As for the amplitudes of significant coefficients, we use multi-threshold to optimize the process of SAQ according to HVS, and consequently enhances the coding strategy.

Biorthogonal multiwavelets are generated from related scaling function vectors via multiresolution analysis. In this paper, we present a formulation for construction of biorthogonal multiwavelets with support [-1,1]. Then we show the biorthogonal multiwavelets of this paper remain the property of symmetry-antisymmetry. Finally, an example is given. The multiplicity of biorthogonal multiwavelets in this paper is any integer r.

In this paper we propose a method of embedding a two-valued image watermark into a still gray image based on wavelet transform. The process of embedding watermark is as follows: At first we divide the original image into some small blocks and implement wavelet transform for every small block. Then we change the two-valued image watermark into 1-deminsion sequence of digital watermark and logically embed this sequence of digital watermark into those wavelet coefficients. In order to enhance the robustness of the watermark we group these small blocks into two sets and embed the same sequence of digital watermark into every set of small blocks. In the process of detecting watermark we synthesize the two sequences of detected watermark so as to get a more accurate sequence of digital watermark. At last by the correlativity of pixel in a image we obtain the final treated image watermark. The experimental results show that this method has not only better invisibility but also the better robustness. Specially it is very robust to various attacks, such as mid-value filter, mean value filter, geometric transformation and so on.

Page segmentation is an important preprocessing step of most document image processing systems, such as filtering, enhancement, compression. Page segmentation in this paper is an segmentation document image into three classes: image, graphics, text. This paper propose a new page segmentation approach, based on morphological wavelets and the markov random fields model. The morphological wavelet is suited to document images containing singularities(edges), for the signal analysis filters in the linear haar wavelet decomposition are linear low pass filters and smooth-out edges. The Markov random fields model provides a good tool for distinguishing between different textures. The new approach provides both lower classification error rates and better visual results, especially where document objects may have convex boundaries and one may be embedded in others.

This is where the abstract should be placed. It should consist of one paragraph giving a concise summary of the material in the article below. Replace the title, authors, and addresses with your own title, authors, and addresses. You may have as many authors and addresses as you like. It is preferable not to use footnotes in the abstract or the title; the acknowledgments of funding bodies etc. are to be placed in a separate section at the end of the text.

Based on Radial Template (RT) and spread spectrum technology, this paper presents a novel digital face image watermarking algorithm that is invariant to rotation, scaling, and translation (RST). Robustness is one of the crucial important issues in watermarking. Robustness against geometric distortion with blind extraction remains especially challenging. This paper utilizes human face images' characteristics, and then proposes a model named RT to locate rotated faces in any orientation. This template is designed to find stable features of center-rotated objects in edge maps. Based on skin detection and edge extraction, our method searches for face-like areas and get their orientations by RT searching. Then the candidates are rotated upright and a spread spectrum watermarking algorithm is used to embed watermark in face images. And this watermark system will be used for main applications: copyright of face image database, identification card's watermark, and portrait-right protection. The experiments demonstrate that the scheme is invariant to rotation and translation except scaling.

Image measurement is a newly rising technology in the domain of the measurement. Use of a non-contact optical technique integrated with computer and image processing technique, which provides high measuring sensitivity and fast data acquisition has been widely recognized as a promising method for the above measuring purposes. On the other hand, the measuring accuracy was extremely restricted by the resolution of CCD camera in image measurement system. For improving the accuracy of the image measuring system, this paper proposed a crossed window automatic edge detection algorithm which has the properties of holding edge, reducing noise and setting threshold automatically in case of non-uniform luminary, and furthermore this paper also proposed a new sub-pixel edge detection algorithm based on polynomial interpolation. The algorithms might improve the measuring resolution of CCD camera in the image measuring system by about 60 times.

Now wavelet transform is frequently used on image compression. And because of a great number of data, in which fingerprint image are saved, the fingerprint image need to compress. Therefore, we use an improved algorithm of Embedded wavelet encoding on fingerprint image compression. This algorithm, named SPIHT (Set Partitioning in Hierarchical Trees) algorithm, can bring faster speed and higher compression rate during the process of compression image. Subsequently, it is very convenient in fingerprint matching with the compression image.

In this paper, we have proposed a new wavelet-domain HMM, called local contextual hidden Markov model (LCHMM) for statistical image modeling. The proposed LCHMM can not only exploit the local statistics but also capture intrascale dependencies of wavelet coefficients. The proposed LCHMM is applied to real image denoising. The simulation results show that the LCHMM can achieve the state-of-the-art image denoising performance at the low computational complexity.

Performance of digital watermarking depends on the parameters of embedding algorithm. Most algorithms cannot give the quantitative relationship between embedding parameters and its performance. This paper proposes a new algorithm of bind digital audio watermarking based on SCS. This algorithm can design the embedding parameters according to the requirement. The paper adds the watermark in wavelet domain and compares its performance with that in time domain against attacks of additive noise, mp3 compression and resample. Computer simulation proves that this algorithm is effective and stable.

Image fusion has important applications in image analysis and image understanding. A novel algorithm for 2-D image fusion is proposed in this paper. This is a feature-based wavelet image fusion method using the multiscale edges information of the source images. The denoising and edge detection method are integrated together for the purpose of obtaining good image edges. The method of selecting optimal decomposition level has an explicit physical meaning. The statistical criteria are used to evaluate multifocus image fusion performance. The entropy and standard deviation of the fusion image is compared with the source images' at the same decomposition level. Experimental results demonstrate the effectiveness of the algorithm which increments entropy and standard deviation of fusion image.

Among all the fingerprint identification/verification systems, the performance relies heavily on the quality of the input fingerprint images. In order to ensure the performance of an automatic fingerprint identification/verification system, it is essential to incorporate a fingerprint enhancement algorithm in the minutiae extraction module. This paper proposes a new method for fingerprint image enhancement with texture filtering, based on the redundant (à trous) wavelet transform, which can much improve the clarity of ridge and valley structures. Experimental results show that the enhancement algorithm is robust for improving the minutiae detection.

This paper shows a new attack method: Aimed attack method in digital image watermarking. This attack method can be used to do vicious attacks, once attackers know the watermark domains and the equations using in the watermarking. Then the author gives a kind of Aimed attack based on wavelet transform. Experiments show that if there is no definite design to resist these attacks, watermark embedded in the data can be destroyed easily.

The Images registration has been a widely studied problem. This paper presents a novel registration algorithm based on wavelet decomposition and mutual information measurement. The feature points are extracted by Simoncelli filter. And a hierarchical feature-matching algorithm is adopted. The transform parameters are gotten by the maximum mutual information as a metric. The simulation results for remote images show that the method can effectively, quickly and accurately work.

Sharp variation points are among the most meaningful features for characterizing transient signals. As a result, The Sharp variation points can be detected by zero-crossing, and they have been selected as the meaningful features for iris recognition. It is less sensitive to the influence of noise and illumination. But the numbers of zero-crossing points is not fixed, whish is unlike the same length of iris feature codes as Daugman proposed, so it will be difficult to match. Boles used the dissimilarity function to solve this question, but it need long time to compute the dissimilarity values. For Huasdorff distance reflects the most mismatch extent of two sets and it will not require the two sets have same length elements, we used the modified partial hausdorff distance to match the feature points. In the experiment, the iris images were corrupted with varying degrees of white Gaussian noise to find the effect about noise. The result showed the method was less sensitive to the change of noise. In additional, we compared the modified partial hausdorff distance matching method with the dissimilarity function method and found that the modified partial hausdorff distance matching method's computing time is much less than that of the dissimilarity function method. Under experiment circumstance, the matching method performed effectively and had a better robust.

Denoising Aerial photos is very important in modern war. Although wavelet-domain hidden Markov models have been successfully applied in aerial photo denoising, the computational complexity of training data is unbearable. The universal hidden Markov tree model that needn't training the observed data is introduced to solve this problem. It is shown that the algorithm based on the uHMT is more efficient than standard HMT algorithm.

In the recent years, with the rapid development of the network multimedia systems, digital watermarking technique has become a feasible solution to copyright protection. In this paper, we propose a watermarking system for still images on basis of scale quantization and coset. The method utilizes a suitable scale factor to scale host image, then according scale version of host image to construct coset for embedding watermark information. Our scheme makes a tradeoff between imperceptibility and robustness, and achieves a higher capability against steganalysis than most of the current quantizing directly coefficients embedding methods. The extraction algorithm retrieves the hidden information bits without host images. Experimental results indicate the robustness of the method to general signal processing attacks.

Objective of image fusion schemes is to extract all the useful information from the source images. Physics-based explanation of image fusion is proposed in this paper. Multiwavelet transform (MWT) offers simultaneous orthogonality, symmetry, compact support, and vanishing moment, which are not possible with scalar wavelet transform. The careful study of multiwavelet transform image fusion has been lacking. In this paper, optical image and IR image are fused using multiwavelets transforms. Image fusion performances of D4 scalar wavelet, GHM multiwavelet, and CL multiwavelet are compared quantitatively. Experimental results indicate that the entropy of fusion image using CL multiwavelet is larger than other fusion images.

Nearest neighbor search is emerging as an important search paradigm in the high-dimensional image databases. The VA-File approach in which data compression technology has been applied can accelerate the nearest neighbor search by reducing the I/O complexity. In this paper, we propose a new VA-File approach based on a multiresolution data structure. Due to this multiresolution characteristic, we can dramatically reduce the computational complexity of VA-File by eliminating improper candidates with much smaller computation at lower levels. This new method has been applied on the k-nearest neighbor search in the image database and has been compared with the VA-File and multiresolution sequential scan method. The experiment results show that the new multiresolution VA-File can improve the search speed and outperforms the other methods.

This article discusses the similarities among wavelet transform, fractal theory and the problems of information storage contained in organism cells' omnipotence, advancing data compression mechanism theory implied in bio-information storage. It develops a new thought to realize high-speed data compression with a high ratio.

This paper proposes a novel digital watermarking method based on singular value decomposition (SVD) with Turbo code enhanced robustness. The watermark, a visually recognizable binary image, was permutated and encrypted by chaotic series and then inserted into the maximal singular value of the cover image's blocked SVD-based transform domain. Turbo coding is used to achieve the robust enhancement in order to make the watermark stronger to resist attacks of common image processing operations and JPEG lossy compressions. Experimental results show that the proposed techniques perform well in both security and robustness.

Firstly, the remote radio monitoring system and the algorithm of JPEG were introduced. Then, the details of the two means by which embed words information into a piece of JPEG image were discussed, such embedding was made in both the forefront and the terminal of the system. The two means were proved feasible by experiments.

With the recent rapid developments in the field of sensing technologies multisensor systems have become a reality in a growing number of applications. The resulting increase in the amount of data available is increasingly being treated by image fusion. Image fusion algorithms provide an effective way of reducing the total amount of information presented without perceptual loss of image quality or content information. In this paper we present a novel approach for fusion of image information using a multiscale image processing technique with a reduced number of levels.

The partial-discharge(PD) signal usually mix up all kinds of interference signals, it has some limitation to extract faint partial-discharge impulse signal from powerful interference signal with traditional analytic method. In this paper, starting with the wavelet transform feature, we have researched the wavelet transform actions in the fields of eliminating white noise in the partial-discharge signal and extracting characteristic parameters from partial-discharge signal. exactitude of this method has been verified by simulating experiment, the result evince that the wavelet transform has very good effect on the fields of eliminating noise and diagnosing partial-discharge failure, so it will be worth studying further.

The transmission and reception of a weak signal in speech frequency under strong pink noise background is discussed. A self-similar signal is generated from a sinusoidal signal by using fractal modulation and transmitted into the channel. As the transmitting signal and the pink noise in the channel are both self-similar, the coefficients of fractal demodulation at the receiving end are just the sinusoidal signal and white noise, using chaotic detection method, the transmitting signal whose signal noise ratio SNR = -80dB can be detected out. This scheme has applicable value in hidden communication and low probability of interception communication.

In low Signal-Noise Ratio (SNR) environment, it is difficult to detect linear FM signals effectively with Spectrogram-Radon transform. To improve the detecting performance, Wavelet transform is applied as a further processing step in this paper to eliminate the noise of the image obtained from Spectrogram-Radon transform. Simulation results demonstrate that it is effective to detect linear FM signals with the method proposed in this paper.

The wavelet transform is widely used for detecting and recognizing partial discharge signal (PDS). This paper introduces wavelet transform theory and application nowadays. By using real wavelet and complex wavelet transform to identify partial discharge signal and restrain disturbance, we gets the advantage and disadvantage of above two wavelet transform methods respectively, the complex wavelet transform (CWT) are more superior to original real wavelet transform(RWT). Simulated applications and the analysis results from laboratory show its superiority in identifying PDS feature from strong disturbance. RWT feature of white noise, periodic noise and PDS is different, white noise and periodic noise can be wiped off with RWT, but RWT only partly extract feature of signal. CWT extracts full- scale feature from both amplitude and phase, it will be widely used to open out tiny feature difference between PDS and other noise

The basic principle of FM-CW radar level gauge is analyzed briefly. A method of estimating its beat frequency is proposed based on periodogram. But the frequency-estimation accuracy is mainly determined by signal-noise-ratio (SNR). So a method of estimating noise level of the beat-frequency signal is then proposed based on Daubechies wavelet. Simulation results show that the noise-estimation method is valid in a wide range of noise background and suitable for an on-line use in FM-CW radar level gauge.

Because of direction space feature of fingerprint image, direction information is important description measure for ridge, which can be used in preprocessing, feature extraction, classification and matching procedure, so it is very important to keep and emphasize intrinsic feature of fingerprint image, reduce pseud-feature point number and guarantee that true feature points be not lost. As a result, in this paper, the approximating algorithm in predefined direction and RAO improvement algorithm are discussed, their weaknesses are analyzed, and a new algorithm is put forward. Based on it, the disadvantages, such as the time complexity is very high, the accuracy is low and they are not suited for fast identification online, almost exist in now matching algorithms, which are analyzed in details. By means of the direction information result, combined with B spline fit, the three order spline curve fit matching algorithm is represented, which has the follows characteristic: when one vertex moves, only the local feature of a piece of curve is affected; it is of the geometric invariance feature that is irrelevant to figure location and coordinate; and it is of changed difference contraction and curve modeling flexibility. Its experiment results give the running time, false accept ratio, false refuse ratio, rotation and translation invariance, at the same time, it can be greatly tolerated that part feature points are lost or distorted, which has verified the validity and correctness of the previous direction information extraction and matching algorithm.

The channels of mobile communication always suffer serious frequency-selective fading. Multicarrier as one of the important modulation technique, has been proposed to cope with it for its advantage in frequency diversity. Traditional multicarrier is characterized by the fact that the time-domain orthogonality is based on the function of sine and cosine□Owing to the fixed frequencies, multicarriers may not perform satisfactorily under frequency-selective fading. This paper adopts a new class of carriers, wavelet- packet orthogonal carriers. The combination of the wavelet waveforms in different nodes in a wavelet-packet full binary tree enables frequency diversity capability and facilitates implementing multi-rate streaming, which retains well time-domain orthogonality as well as to provide subband diversity to mitigate frequency-selective fading. A new result on the probability density function of the sum of square is proposed. So a closed-form expression of the bit error rate of this multicarrier CDMA system is derived, where the correlated Rayleigh fading channel with an additive white Gaussian noise is employed. The simulation results reveal that correlation among channels decreases the performance of the system. Wavelet-packet multicarrier modulation performs well under same fading situation than general MC-CDMA.

Features of radar signal in a pulse embody subtle features of radar signal. The analysis and extraction of radar signal in a pulse are the hotspot problem and have been deeply studied in the field of electronic war. This paper presents a method, which uses wavelet analysis for features of radar signal in a pulse. The feasibility of the method is studied. Some simulated radar signals are investigated using wavelet transform and results indicate that the wavelet transform is the power tool for features of radar signal in a pulse.

A novel technique for the secret video communicating based on the discrete wavelet transform(DWT) is present. Transform a intra frame of video to three gray image firstly, and then compute the Lth-level discrete wavelet decomposition of the gray images, with which the secret data W is embed simultaneously and do invert wavelet transform to obtain the gray images which contains the secret information. Change the intra frame of video based the three gray images to make the intra frame contain the secret information. When extract the secret information, transform the intra frame to three gray image and do Lth-level discrete wavelet transform to the gray images, distill the secret data W' from the wavelet coefficients of the three gray images. Test results show the superior performance of the technique and potential for the secret communicating of video.

The each layer's productivity of the oil well producing many layers joined together is usually made certain by the means of production logging, each layer's testing and tracking by tracer, but their costs are high and periods are long, and cann't be applied on the well of cross layer or low velocity of flow. Sometimes, need shut well and stop production, these busyworks are easy to harm oil layers. Adopting the Crude Oil Chromatogram Fingerprint Technology can monitor oil-layer's productivity preferably. Analysis the crude oil fingerprint characteristics of the different layer, make sure the fingerprint differences of these crude oils and reasons of bringing the differences, and develop the prescription experiments of crude oils of two layers and three layers, establish the mathematic models of the each layer's productivity to compute, and apply wavelet pack analyzing technique to increase the crude oil fingerprint calculation's accuracy. The computing result indicate that for two layers or three layers joined extraction, the error margins of the actual prescription and the fingerprint Calculation are within 10%, It shows that the crude oil's fingerprint can be used to compute the productivity contributions of each layer in the oil-well with many layers joined to produce, and monitor the productivity's variety of each layer.

In this paper, the gear vibration analysis was performed between common involute gear (CIG) and the double involute gear (DIG) in the same gear parameters and running condition. The FFT and wavelet methods have been introduced and applied to gear vibration analysis. The experiment results show a clear advantage of DIG with the help of Spectrum analysis.

The wavelet transform can decompose a signal into several frequency bands with the multiresolution analysis. Employing this property in heart rate variability (HRV) analysis can calculate some frequency-domain parameters for clinical diagnosis. The foundation of this method is Parseval theorem. The wavelet base selection is the difficulty in this method. To compare the effect of different wavelet base, a series of daubechies orthogonal wavelet bases from db1 (Haar) to db30 are selected for HRV decomposition. A further study shows the quantitative results mainly depend on the amplitude frequency responses of the corresponding wavelet filter banks, the quantitative results gradually converge with the increase of the db wavelet order.

This paper deals with feature extraction in combination of wavelet decomposition with neural networks, under the condition of low SNR. The signal is first decomposed by wavelet transform, and on the base, the decomposed coefficients are reconstructed to form a new time series, from which some energy parameters can be extracted by time-domain analysis. By means of BP network, it is possible to recognize whether target signal is involved or not in received signals. The effectiveness of the method is verified by a real target signal with additive simulated noise signal, especially under the condition of low SNR.

This paper puts forward a novel digital signal processing (DSP) framework for ultrasonic nondestructive material evaluation. The kernel of this framework adopts a uniform wavelet package algorithm architecture based on lifting scheme. Specially, the Joint Time-Frequency Analysis (JTFA) technology, such as Reassigned Gabor Spectrogram, is fused into the DSP framework for the extraction of feature information. The framework is applicable for single-channel and multi-channel ultrasonic testing systems, including phased-array ultrasonic imaging systems.

Radar detects targets by emitting electromagnetic energy, then processing the signal from the reflected radiation. The radar observation of reflected signal is corrupted by additive noise and clutter. It is very difficult to detect weak targets in heavy sea clutter. It has been shown that wavelets transform (WT) can remove noise more effectively for a variety of applications. In this paper we apply the discrete wavelet transform, the Haar wavelet and hard threshold filter to propose the method to detect weak targets in sea clutter. The simulated results are also given and show that the method is very effective.

The application of wavelet transformation to edge detection is described. Based on the wavelet theory and according to a multi-scale edge detection algorithm, a method of transforming two-dimensional edge detection problem into one-dimensional wavelet transformation is advanced.

In this paper, a neural networks, which with genetic algorithm to train its weights, is used in fit the curve of roll shape. At first, the noise of output of charge-coupled device (CCD) is filtered, then it is binaryzation processed, finally the neural network, which with genetic algorithm to train its weights, is used to fit the curve of roll shape. By comparison it shows higher measuring accuracy.

Synthetic Aperture Radar (SAR) has the capability to map the ground with high resolution and is used in both civilian and military applications. Moving targets with respect to the stationary background cause smeared and ill-positioned images. Detection and imaging of moving targets in SAR signals are necessary. We propose a method combined wavelet transform (WT), ridges of the WT modulus and Radon transform (RT) to detect moving targets. Processing steps of this method and simulated results are given in detail. Simulation results show this method can detect moving targets and estimate its parameters in noise case effectively.

This paper introduces one way which based on wavelet analysis to assist people to learn English words more effectively. It imitates the method that we study Chinese words when we were very young. Moreover, the process also be given to test the accuracy. Finally, we put forward a experiment to support our proposition and get satisfied result.

In this paper, a palmprint, which mainly contains principal lines, wrinkles and ridges, is regarded as a texture. A palmprint texture has intrinsic multi-resolution characteristic in different directions. We analyze palmprint textures using wavelet packet and define a novel palmprint feature, called wavelet packet feature (WPF), for personal authentication. WPF can reflect the characteristic of a palmprint texture at different resolutions and in different directions, thus it can efficiently describe the palmprint. In the verification tests, 98.5% accuracy has been obtained by the proposed approach with 4-level WPF. The experimental results demonstrate the power of the proposed approach.

A general method for the construction of nonseparable orthogonal scaling functions with dilation M = aI_n, where a ≥ 2, and a ∈ Z is presented. The wavelets can obtain a more flexible tiling of the time-scale space than in the case a = 2. What is more, the approach of this paper can be generalized to construct wavelets related to a general dilation matrix.

Wavelet analysis is internationally recognized up to the minute tools for analyzing time-frequency. It is chiefly due to the "adaptive feature" and "mathematical microtelescope feature", wavelet analysis is becoming a focus point of many sciences, and is fondly delighted as tools for so many scientific workers. It plays an important role in the signal & information processing. In clothing system the temperature and moisture difference between human body and environment results in constant exchange of heat and moisture. This paper puts forward the ideas of applying the wavelet theory to analyze & process temperature and relative humidity signals of clothing system by measuring the changing course of skin temperature, temperature and relative humidity in clothing. It is expected to find out the regulation and establish the pattern of the heat-moisture transmission of clothing system.

Firstly, Some theorems that 1-D projections of a 2-D mother wavelet are 1-D mother wavelets were proved. These theorems are useful to detection of edges and filtering of noise from projection data of CT (computerized tomography). Then, we give some experiment results.

Applied the discrete wavelet transform to extreme –value theory, a new estimator for the extreme-value index is conceived with scale function and samples. We prove its weak consistency under the condition of regularly varying and the asymptotic normal under the condition of second-order regularly varying.

It is always a hot subject on research and produce of single-phase self-excited asynchronous generator. The focal points are the unstable output voltage. Theoretical research and practical experiment are developed on this question in this paper.

Web Intelligence (WI) presents excellent opportunities and challenges for the research and development of new generation of Web-based information processing technology, as well as for exploiting Web-empowered advanced applications. The paper investigates how to develop intelligent portals that enable e-business intelligence and deal with the scalability and complexity of real world, efficiently and effectively, by using WI technologies.

By introducing the notion of Information Repository and evolving original mereological relationship among information granules, the Granular Representation Calculus (GRC) employed as fundamental modeling mechanism is extended to represent Internet information sources, including the multimedia resources described in Resource Description Framework (RDF). With information granules represented in terms of the new versioned GRC, a framework of Ontology-Driven Web Information System is briefly introduced as a functional prototype of new generation paradigm for media system.

In order to supervise the driving behavior effectively and regulate safety steering, focusing on velocity and interval between vehicle during driving which are the two most important factors related to safety. A knowledge discovering method to discover the driver's steering habits and regular pattern is presented from the driving record, with the method of KDD (Knowledge Discovery in Databases), which can examine steering safety and improve the driver' safety consciousness greatly.

A novel ant colony algorithm based on hybrid behavior (ACA-HB) is proposed in this paper. First three ant's behaviors are defined in ACA-HB and then the ant colony is divided into three sub-colonies, ants in a sub-colony perform same behavior. When moving from city to city, ants in different sub-colony use different transition formulas (perform different action), the diversity of the algorithm is greatly controlled and stagnation behavior is avoided in this way. This is greatly different from common ant colony algorithm. Comparison between ACA-HB and ant system (AS) shows that ACA-HB not only avoids stagnation behavior but also outperform AS.

In this paper, the sandwich sets in semigroups are discussed. It is found and proved that there are some classes of semigroups in which every sandwich set has exact one element. Furthermore, the sandwich sets in lett C-semigroups are also discussed and a special case of the sandwich sets is presented.

Studies and puts forward an information manufacturing platform prototype specifically for the mechanical product-oriented design. Based on the IT, area networks, and mechanical design and manufacturing technology, and centered with the mechanical product design, the platform consists of three levels. Fabrication level of the mechanical product, product design level, and components design level. Within a manufacturer the three levels are connected with IP, forming an area network, extending to other relevant manufacturers. The core of the platform, mechanical product design level, relying upon CAD/CAM/CAPP/CAGS, asks the higher level for, or receives, design instructions, and reports or provides information and data for the components design level. The stress is given to the structural members and the information access channel construction. All these three levels may, in order, or directly, get in touch with the manufacturing planning and decision-making level. An example of prototype of a lathe spindle component is given by using the platform.

Based on an proposed operation named S-multiplication and the intractability of discrete logarithm of finite field GF(p³), a new verifiable multi-secret sharing scheme is presented in this paper. Compared with the existing secret sharing schemes based on exponentiation operation, this scheme has an advantage in computation efficiency, and can be used to share multiple secrets for a cooperative group in certain communication environments.

A new modeling method of classification engine based on unidentified short sequences of system calls is proposed. Mobile agents can carry the classification engine easily in the distributed multi-agent intrusion detection systems. The classification engine has the advantages of the concise pattern set, high accuracy, fast transportation and easy comprehension. The core spirit of the modeling method is that the local behavior characteristic hidden in the unidentified short sequences of system calls can be exactly distinguished on the statistic meaning with the help of adjacent correlative information. Experiment results indicate that the classification engine clearly improves detection accuracy against known and unknown attacks.

Research of early fault intelligent diagnosis for large scale machine equipments has great social and economical significance. It directly decides production safety and economy benefit of production units. If big equipment stopped working for one day, the loss would be much! The coal excavator, which is one of most important equipments used in mine, will be of thirty thousand dollar loss if stopping one hour. So it is very important to guarantee coal excavator form faults and make it work without errors in all cutting process, that is base on earlier diagnosis and earlier maintain of fault.

But it is not ease for faults to be earlier diagnosed, which needs actual-time measuring and testing of work characteristic of coal excavator and understanding of deep level knowledge of its model. We apply expert system and neural network to catch all kinds of information from sensors, extract feature from them to obtain accurate message of coal excavator in running process and finally carry out information fusion to achieve result of early fault intelligent diagnosis for coal excavator. Although it is designed for coal excavator, it may be used in other fields, too.

Embedded technique based on Internet will provide the more convenient means of remote surveillance and control. This paper makes use of the characteristics of Ethernet to design the embedded temperature controller. Web page, temperature display and upper or lower limit setup of the temperature are realized by Dynamic C. The temperature can be controlled and monitored by Internet. TCP/IP protocol, HTTP are involved in. The introduction in this paper has the significance for the study of the embedded control system based on Internet.

In this paper, we present a novel technique of intelligent agent, which can perform digital watermarking scheme on variant file types such as jpeg, bmp, wav, mpeg, etc. "Watermarks" embedded in are XML self-describing information of corresponding files, which are invisible and can be transferred with their related files when they are copied or moved to different places. An agent can "understand" a file by extracting the self-describing information embedded in it. With such information some critical requirements can be satisfied, like finding a picture with more than three beautiful girls or calculating how many songs sing by Chinese children existed on local network. Finally, scalability and efficiency of the agent are discussed.

As a new computation paradigm, the Mobile Agent (MA) paradigm seems to be a promising technology for developing applications in open, distributed and heterogeneous environments, such as the Internet. Although security is very important for mobile agent system, security has not been pursued in a satisfactory way yet. After discussing the security problems and the existing security technology of mobile agent, a dynamic security adaptation mechanism is presented. Using this mechanism, different security implementations can replace one another base on given security policy dynamically. This mechanism provides a framework with which mobile agents can select security implementation dynamically from different security algorithms implementations to adapt different applications and run time environments on demand. We also draw a distinction between adaptation and non-adaptation. Our solution for dynamic security adaptation provides a concept for exchanging environment dependent security implementation of mobile agents dynamically during runtime. Dynamic security adaptation enhances efficiency of mobile agent in terms of bandwidth and scalability on the premise of ensuring the security of itself.

Middleware is essential to migrating mainframe applications to client/server applications and to providing for communication across heterogeneous platforms. Communication between the client and server is always the emphases of research of middleware. So far, the primary mechanism of it is still RPC and OORPC. This paper presents the design, the implementation, and the evaluation of a novel mobile agent-based middleware. Our architecture use mobile agent to replace RPC and OORPC to resolve the disadvantage of traditional middleware.

Over years, a lot of people have devoted themselves to the research of parallel I/O. In some applications, Parallel I/O, whose goal is to improve the performance of a system by enhancing its throughput capacity and accomplished by reducing bottlenecks in I/O path: no single I/O device, no single I/O bus, and no single network path, has been used. Nowadays, various web services have very high requirement of throughput. In this paper, we discuss a parallel I/O architecture and describe how to improve the performance of web service with the parallel I/O system which is at the back of database sever in the 3-tiers architecture and provides transparent data access. To provide parallel I/O, three software requirements have to be satisfied: application interface, underlying high performance data storage mechanism, and tools for file manipulation. And as the parallel I/O system is used in web applications, parallel I/O will interact with database server.

In this paper, the characteristic of the nonlinear complex giant system in the information age war is analyzed and the new requirements of modeling and simulation are discussed first. Then, on the base of analysis of multi-agent system, the creative methods of military modeling and simulation and the corresponding model system is studied. Finally, the question of traditional modeling and simulation can't fit the antagonism between system in the information age war simulation is solved.

Agent technology is a good technology in the development of Multi-Model Processing and Intelligent Expression systems. Iit is a best choice to use agent technology in the realization of Multi-Model Spatial Analysis and Intelligent Expression. In this article, we firstly introduced the agent technology and the significance to realization Multi-Model Spatial Analysis and Intelligent Expression. Then the necessities for agent technology to realization Multi-Model Spatial Analysis and Intelligent Expression were explained. The architecture of Agent-based Multi-Model Spatial Analysis and Intelligent Expression we put forward was introduced in the next part. Next, we expatiated the realization of Agent-based Multi-Model Spatial Analysis and Intelligent Expression using java technology and the applications in the Geographic Information Service Platform of "Digital Fujian" in China. The application system has validated the feasibility of our research. At the end of this paper, a conclusion of this paper is also discussed.

Information classification and encoding is the most important job for the information standardization, until now we have no national standard for geographic information classification and encoding which is obstacle for the information sharing and service. How to use the established domain specifications of classification to form a national standard and solve the conflict and overlap between them is a big problem. Using ontology, we can easily resolve the problems the former classification scheme have and unite them together, realize the updating of the classification with the knowledge changing, simplify the classification and encoding process, realize the reusing and interoperating of the meta-entities, and in the end help to realize the information sharing and service in the network.

This paper proposes a modified constant modulus algorithm for blind equalization of time-varying channel. The new algorithms replace constant step-size by the variable step-size that was controlled by the nonlinear transform of residual error. The advantage of the new algorithm is that it has the lower residual error and the faster convergence rate than the constant modulus algorithm. The superior performance of the proposed algorithm is illustrated by the mean square error curve of 8-PAM signal transmitted through time-varying channel. The simulation results demonstrate that an improvement in performance is achieved with the proposed equalization algorithm.

Due to the combination of embedded technique and Internet technique, it is possible to control field equipments through Internet. This paper presents that the embedded controller can control analog output through web browser. It is of great significance. Ftp server which is the network application of embedded controller is also studied in detail. Network connection, IP address, Dynamic C and network program are illustrated completely. The exact example of network application is given out. The introduction in this paper has the importance for the study of the network application to the embedded control system.

The high-speed and real-time data acquiring system based on VXI Bus is introduced, including hardware structure and software structure, its emphasis is introducing the software design of the data acquiring system.

This paper describes a shift calling computer network system in the railman apartment, which includes the upper computer (PC) and the lower computers (singlechips) and the communication between these computers. The system is implemented with advanced software platform and is of complete functionality and practicability.

Discussed the advantage of Anti-Radar hidden coat in the future's development on the basis of the contrast of two kinds of Anti-Rader hidden technology approach: Making use of the chebyshev polynome, nonlinear equations are established. Based on CAD, the simulation result is applicable to the demands that the reference 2 and 3 put forward.

This is the first paper on the relationship between distributivity and Zadah's operations. When T-conorm and T-norm is Zadah's "intersection" and "union" operation respectively, pseudo-addition and pseudo-multiplication is Zadah's "intersection" and "union" operations respectively, the sufficiency and necessary conditionson them can be obtained. The theorems in semiring can be applied.

In recent years, strong and challenging requirements of modern society are involved in advanced surveillance systems, which aims to increase safety and security. At the same time, fast improvements in network technology, multimedia technology, and computer science make it necessary to consider new perspectives in this field. Today, the research of P2P Computing, Media Streaming and Adaptive Service has captured the interest of both the academia and industrial worlds. In this paper, we proposed a P2P based adaptive service framework for the multimedia surveillance system. In particular, aspects related to the applicability of P2P to the problem of multimedia surveillance system are discussed and a new surveillance system structure is presented, which bases on P2P model and uses media streaming system structure for reference. The surveillance software architecture is designed, which bases on Adaptive Service architecture and uses COM technology. To be adapted to the available systems resource dynamically, the adaptive control and manage strategy are introduced. The advantages of these solutions are as follows: 1) increased flexibility and configurability, 2) optimal use of available system resources.

This paper introduces two experiments practicing for digital Asia, the technical Intercultural Collaboration Experiment and the non-technical China-Japan-Korea Virtual Community Experiment. By analyzing the two experiments, we found that, it is possible to use machine translation technologies to make the effective communication among people in Asian countries. The analysis also shows that human interferences and adaptations to machine translation is a solution to break the language barriers in multilingual collaboration and intercultural communication.

In distributed Geographic Information Systems (GIS) and Web-GIs, remote access to geospatial database is recognized as a time-consuming task. Progressive transmission, a technology dealing with massive amounts of data and slow communication links, can be used to solve this kind of problem because it allows users to start work with a partially delivered dataset. In this paper, we particularly discussed this problem and propose a method, by which we can progressively transmit the sequence of levels of detail (LODs) of map representation while preserving topological consistency between different levels. The polyline simplification is not accounted suitable for this type of task because it is time-consuming and may cause topological error between origin object and its simplification. In order to overcome the defect, we employ the algorithm of minimum number of safe sets (min-SSs) to simplify the polyline. In addition, we propose a structure based on Simplicia1 Multi-Complex (SMC) and Reactive-tree to avoid data redundancy as well as reduce the space complexity of geospatial database. We show that the volume of coarsest level data can generally be reduced to one tenth of that of the finest level, and employing appropriate polyline simplification can speed up the transmission of vector map data.

The present paper proposed the system structure of the intrusion inveiglement, and designed the prototype system of the intrusion inveiglement, was to an existing network security system important expansion. Through aims at the tool, the strategy as well as the attack goal which the intruder uses carries on the analysis, to guards against well intrudes and attacks has the certain theory instruction significance.

Combined with the practice of developing an official automatic management system, database management system meeting user's requirement is designed based on network and database technology, and the implement scheme is analyzed.

With the development of the Internet and computer technology, the cost, size, and power requirement of microprocessors have drop to a point [1]. It is possible for us to merge the capabilities of computer with the capabilities of Internet to expand their area of applications. The development of technology can provide us with high performance MCU, and it is possible to design an embedded computer with its operating system to connect with Internet. For integration aspects using Internet services and embedded systems can provide a suitable solution to overcome certain critical issues in exploiting remote facilities. It is the inevitably trend in most of network projects in the future. In this paper we present an approach to use Internet and embedded systems for remote medical care, and give an introduction of the structure, of the constitution of the embedded computer developed by our team, of the operating system Vxworks, at last show an experimental result made by us.

To the VR fuel depot fire fighting training system, the architecture of the VR system is hard to build. According to the classification of the VR scene, the corresponding multi-agent system has been put forward. The reactive agents have been used to set up the 3D scene and simple behavior entities; the deliberative agents are agreed with the need of intelligent entities. The application shows that the multi-agent VR training system can primely solve the difficulties of traditional training methods.

No abstract received.

This paper presents a collaborative filtering recommendation system combined with market value functions. In the approach, by building market value function model for each item, we can produce a user rating information table. Then Collaborative Filtering algorithm can produce a list of recommended items. Experiment in real world examples shows that the method is effective.

For the video surveillance system, the most important thing that it cares about is changed images. It's necessary to research on the robust and real-time movement detection and abnormity alert algorithm. In this paper, a multilevel determination based movement detection algorithm is proposed. It is used to eliminate the influence of vibrating pixels, simultaneously guaranteeing the sensibility of real-time detection and the adequacy of long period of saving. The experiments show that the algorithm is robust to noise and vibrating pixels, and it can performs on real-time.

This paper presents results on appearance-based three-dimensional object recognition (3DOR) accomplished by utilizing a neural network architecture developed based on Kernel Principal Component Analysis (KPCA). The basic idea of KPCA is first map the input space into a feature space via nonlinear mapping and then compute the principal component in the feature space. In this paper, we are utilizing the KPCA technique to enhance the object recognition. Through adopting a polynomial kernel, the principal component can be computed in the space spanned by high-order correlations of input pixels. The imitate results confirmed the effectiveness of the proposed method.

A new surface simplification algorithm based on vertex degree and costing increment for producing high quality approximation of polygonal models rapidly is developed. By introducing vertex degree and costing increment into Iterative Edge Collapse (IEC) algorithm, the new algorithm can reduce computation when measuring an accurate geometric error on a highly curved and thin region, and can preserve key points steady so as to maintain some iinportant features. Two examples are employed to test the validity of the surface simplification based on vertex degree and costing increment algorithm.

In this article, two sections of pre-process of filtering wave and making thinner are discussed. By using of KARLMAN filtering wave algorithm, images with colorful noise are wave-filtered. By elimination shell algorithm to make thin of image, the line description is made.

Considered change of web service manner, a new performance analysis model for web server system is proposed, and program average length f and dynamic service manner of web server are analyzed. Moreover, SPEC web99—the web server performance test instrument is used to discuss the effects to system performance. It has practical meaning for network managing, planning, and upgrading.

Among many kinds of IDSs(Intrusion Detection system),how to recognize and identify network intrusions is a base criterion to evaluate an IDS. Traditional IDS implements this by security rules matching. However, new network intrusions emerge in endlessly. Obviously this detecting way can't meet new situation' s need and will be out of date in the near future. Aiming at this problem , This paper bring forward a new kind of recognition model of network intrusions which make use of immune antibody's self-organized learning ability and may accomplish the recognition for variation of network intrusions properly.

Electromagnetic waves in GIS excited by partial discharge (PD) propagate not only in the transverse electric and magnetic mode(TEM), but also in transverse electric (TE) and transverse magnetic (TM). This paper deduces out the expressions transient field excited by PD in GIS anywhere by using dyadic Green function theory, and the contributions of TEM mode and all high order modes in the detecting frequency scope should be taken into consideration when calculating the electric field excited by PD in GIS. Inner couplers for the GIS model are developed and their frequency responses are analyzed in antenna model. At last the PD pulses excited by man-made defect are detected through inner couplers using high parameters oscillograph.

In automatic recognition of license plate, character divided plays a very important role. The arithmetic researches and comparisons as well as synthetics are done in this paper. Such arithmetic includes proportional projection, profile projection and fixed space between characters (fixed SBC). A synthetic approach is used to divide the characters in the automatic recognition system of license plate. According to the practical results, the accuracy of character divided can reach 98.7%.

Based on ant system, the scheduling model developing method is discussed in IC assembling and testing application system environment. Under the guidance of general requirement, the proposed methodology is applied to the scheduling of working procedure, a model established with ant system, and simulation result is reported.

The correlation properties of zero-mean chaotic sequences generated by the Tent map in the discrete time domain are discussed in this paper. It is found that the chaotic sequences are nearly orthogonal to both the standard images and their DCT coefficients. The chaotic sequence has a broadband spectrum and the auto-correlation of these sequences is a delta-function and their cross-correlation is approximately zero. As an application, a chaotic watermarking scheme for copyright protection of digital images is presented. In this scheme, the watermark signal generated by the chaotic system provides sufficient watermark complexity. Watermark detection, which can be reliably extracted without referring to the original image, is performed by the correlation of the DCT coefficients of the watermarked image with the chaotic sequences. The experimental results show the robustness of the proposed scheme under JPEG compression, filtering, twirling, cropping and additive Gaussian noise.

Region of interest (ROI) image coding allows for encoding the ROI with better quality than background (BG). In this paper we present a ROI coding method based on multiple bitplanes interleaving (MBI). This method supports arbitrary ROI shape without shape coding. Compared with the general scaling based method, the maxshift method and the BbBShift method, the MBI method has more flexibility to control the reconstructed image quality of the ROI and BG. The complexity of the method is kept low.

Index terms: Image coding, region of interest (ROI), bitplane coding.

This paper presents an improved Local Energy model called Symmetry Local Energy model (SLE), which is a novel image feature detector based on phase information. SLE extracts the image features with only four symmetric phases and is calculated based on log Gabor wavelet. The results are consistent with the characteristics of human visual systems. Not only can SLE detect the luminance profile, such as step and impulse discontinuities, but also can detect the features in Mach bands and in sinusoidal grating at the same time. These cannot be done simply by looking for local maxima in intensity gradient or by looking for local energy maxima. Compared with local energy model (LE), SLE can also improve the anti-noise ability and localization accuracy.

On the basis of analysis and research, we built the concept of super-resolution processing to blind Image, presented that estimate degenerative function (or degrade definition function). firstly, we estimate the best degeneration function by repeating operation, then extract the best reconstruction image, on the case of only have observation image and haven't any foregone knowledge, or by the several best estimated degeneration functions, estimate several reconstruction image, then deal with the data together. By simulation, it declare that the method is well, it provide a new way to super-resolution processing on single image.

It is important that how to realize the sequence control. Sequence control function can realize the intelligent conjunction. This paper analyses the sequence control example of the distributed control system. It is of the significance that various types of functions including regulatory control and sequence control can be combined together. Many functions, such as the timer, software counter block, PID controller, local switch, common switch, the relational expression block and operated guide messages, are involved in the realization of the application illustration. The usage of the step label for sequence table is described in detail. The configuration of the control project and control drawing are studied in particular.

Based on the analyses of GML implementations and current spatial data engine technology, this paper proposes a distributed spatial data engine model that deals with operating almost all spatial-relevant objects in large object-relational databases. Components and organization in this model, together with its communication mechanisms, are all based on XML technology. The paper gives a detailed explanation of this model, including its architecture and key technologies used.

Using geographic information through internet all over the society is the tendency of the development of GIS related theory and technology, and the concept of "Geographic Information Service" was born in this environment. In this paper we firstly introduce the necessity and objective of construction distributed "geographic information service system" for Digital Province in China. Then based on the analysis of "Geographic Information Service" stated by ISO and OGC, we put forward the architecture of the "Geographic Information Service" for Digital Province in China. In the end, aiming at the requirement of building "Digital Province", the realization case to build the "Geographic Information Service System" based on internet is discussed in detail. The experience in building "Digital Fujian" is introduced and the development direction is also discussed.

To satisfy the requirement of embedded video encoder in various scenarios, a design scheme of Embedded Video Encoder Based on DSP (EVC) is proposed in this paper. To verify the feasibility of the EVC scheme, TMSC6202 DSP of TI Corp is chosen to construct a hardware platform. By studying efficient video coding technologies and system optimizing implementation on the platform, a high performance multi-mode MPEG-2 video encoder is achieved. Experimental results of the video encoder show that the EVC scheme is not only feasible but also general and flexible, thereby a new video encoder scheme is provided to replace previous ones based on PC and standard ASIC.

Personalization is a major characteristic of Web intelligent system for the future. To provide personalized commodity recommendation, this paper presents a new recommendation mechanism based on multilevel customer model, which formalizes the recommendation process as the learning of multilevel customer model, the generating of recommendation set and personalized filter. The mechanism captures customer's needs from three aspects: shopping demands, preference characteristics and purchasing power, thus enhances the degree of personalization in recommendation and improves the effectiveness of recommendation.

Collision detection acts as an important role in Virtual Reality Environment (VRE), the Layered Bounding Box (LBB) algorithm can effectively resolve the inherent complexity of computing in collision detection. With the number of objects increasing, the burden of the collision detection will dramatically increase in VRE. This paper discusses the LBB algorithm and proposes an improved object-oriented algorithm in VRE, which can greatly reduce the computation costs of collision detection in VRE system with no losing of the accuracy.

In this paper, a new class of completely generalized nonlinear fuzzy implicit quasi-variational inclusions involving generalized m – accretive mappings are introdued and studied. By using resolvent operator technique for generalized m – accretive mapping in q – uniformly smooth Banach space, an existence theorem of solutions for this kind of variational inclusion is established. A new iterative algorithm is suggested and analysed. The convergence of the iterative sequence generated by the algorithm is also given.

To turn automation diesel generator set into Power station without person's joining, PLC was born. Taking a set for example, the detailed explanation of PLC's course in apply has been accomplished.

With the fast development of modern technology, automation becomes more advanced and advanced. People's labor is replaced gradually. As a result, automatic diesel generator sets which save workforce have been used as supplies in emergency, especially in telecom department, electrical plants, hospitals, schools, hotel and so on. Without people's care, intelligent diesel generator sets can automatically start, adjust speed when main power is abnormal. It also can switch on and load. Furthermore, it can automatically switch off and stop when main power is normal. As is well known, Programmable Logical Controller play a important role in intelligent control. PLC has good performance at the low cost, small shape and high speed. With many inside counter, timer, I/O relays, memorizers, PLC may be programmed with logical language. Thanks to PLC, the diesel generation sets system is more flexible and reliable.

Based on the idea of Stage-Gate System (SGS), the life-circle of Virtual Enterprises (VE) can be divided into several key stages. The important issues are the information's mining, interaction and controlling. Clustering is the one of important technology in data mining. In order to improve the efficiency and quality, in this paper, we adopt pretreating technology named data partition before clustering. After that, we can cluster parallel and improve efficiency. Finally, brief conclusions are given.

In production and exploration of oil field, hydraulic fracturing is an important method to increase crude oil recovery ratio. Many large petroleum corporations and service corporations applied themselves to the research of hydraulic fracturing and developed many software which can't fit to applications at the present time, So it is urgent to use new technologies such as web database technology to development new systems to fit with the need of applications of now. This paper use an sample to illustrate the application of web database technology in hydraulic fracturing of petroleum industry and put forward some opinions and resolvents about issues which may appear during the applications of web database technology in petroleum industry.

Symmetry exists widely in the real world and plays a remarkable role in perception problems. The existing symmetry detection algorithms are mainly gradient-based. Instead of using gradient information, a phase-based symmetry detection (PSD) algorithm is proposed. PSD is calculated based on log Gabor wavelet. The experiments show that this algorithm can be applied directly to original images without any pre-knowledge and any pre-processing, such as segmentation; it is invariant to rotation, luminance and contrast; it can detect several types of symmetries at the same time, such as rotational symmetry, bilateral symmetry and curve symmetry.

Currently there are various data processing methods in power quality monitoring system, and this paper put forward a new data processing method based on data pre-processing, reasonable choice of time interval for analysis and generating daily statistical file. According to IEC international standard, it is effective to get clear and explicit graphs and quickly generate different statistical reports in control center by adopting the effective measuring time in collecting apparatus and reading the data file of monitoring spot in accordance with appointed format in monitoring system. With this method of data processing, the whole monitoring system of power quality can be effectively applied to practical power system.

An online self-adaptive decision-making system is described. The system emphasizes the team's strengths while exploiting the opponent's weaknesses. The problem is discomposed into three subtasks: gathering information responding to the opponent; representing team behavior of the opponent and taking tactic responding to the opponent. The respective algorithm is also given out.

Advances in echocardiographic systems and computer applications have made three dimensional reconstruction of anatomical structures possible which open a new and fascinating field of color Doppler flow image, but the spatial shape and quantitative evaluation of mitral regurgitation is often difficult in the clinical setting. In this paper, we present a method to complete the 3D reconstruction of the mitral regurgitation, mitral regurgitation information was first derived from color Doppler flow images, and then the mitral regurgitation velocity values was mapped according to the color bar in the images. With the proper method of interpolation and rendering, the experiment result of 3D visualization of mitral regurgitation is satisfying, Measurements from a 3D reconstructed flow convergence region may be superior to measurements from 2D color Doppler recordings to calculate volume tlow, 3D mitral regurgitation reconstruction is possible and opens new possibilities in flow quantification, Further study of the method may be very helpful in the diagnosis of heart diseases.

The change of brain electrical impedance has much to do with brain Oedema and brain Infarct. This paper has put forward a kind of new method in examining brain Oedema aroused by the shortness of blood, which is called brain electrical impedance contour image(BEICI). Through the method, the value of brain electrical impedance can be used effectively to make an orientation where the block focus is on the brain Infarct sufferer. Moreover, the status of encephalic haematoma and Oedema's development happened in different time could be displayed. The principle of brain electrical impedance contour image, The network Structure and the result tested in clinic together with the analysis are introduced in the paper.

In this paper we present a new approach to extract video objects from video sequence by combining of color, motion information and orientation information measure. The key idea is to use the orientation information measure and color information to find regions with homogenous characteristics and then use motion difference of image sequence to extract the moving objects. The algorithm can make better use of available computer resources to achieve optimal performance with a new multi-resolution technique, in which the low resolution image is segmented by orientation information measure firstly, then sub regions are merged by color information and motion difference, the calculation in the high resolution will only be performed in the boundary of the sub-region, to significantly accelerate the computation. Finally, To obtain the accurate region boundaries, the information of color image edge detection is used.

Enhancement is a crucial step in automatic fingerprint recognition. But most enhancement approaches, performing band-pass filtering, may create spurious ridge information in enhancement. We examine this problem and suggest an adaptive oriented filter as a solution. The experiment shows that our filter is both effective and robust.

In the geophysical exploration on Chinese continent scientific drilling, the gravity anomalies map in Dabie was made and it was analyzed with wavelet transform. The 1-4 order wavelet component and the fourth order wavelet approximation was got. Result we have had is important to divide the structure units and identify recent mantle configurations.

A novel steganography algorithm operating in the wavelet domain is presented. The proposed algorithm gives two kinds of secret messages bits, which can be classified based on their important, and then have these bits hidden in intermediate or high frequency. 0 or 1 bist can be embedded through regulating intermediate or high frequency coefficients. The most important secret messages is hidden in intermediate frequency. which can resist normal JPEG compressing attack, The other is hidden in high frequency, which has high capacity advantage. The simulation results demonstrate stego-image possesses good visual effect after datum is hidden. Through blind extracting, rates of restoring datum are preferably.

Organic pollutants have become widespread pollutants in the water and now represent a global contamination problem. In our present study, green tea extract (GTE) protect primary cultured rat hepatocytes (PCRH) against DNA damage induced by multiple organic pollutants (MOP) in tap water was test in comet assay, and inhibition the chromosome aberration induced by MOP using mouse micronucleus test. Meanwhile, the quantity of malondiadehyde (MDA) in the culture medium released by PRCH was measured. Results showed GTE could protect PRCH against DNA damage from MOP in a dose-dependent manner and inhibit the chromosome damage induced by MOP. One of the mechanisms of protection against DNA damage is inhibiting the oxidizing damage.

This paper presents a creative opinion that analysis of wavelet and amphibolous mathematics can be applied in anthropology research especially in Youyang, Xiushang, Laifeng, Longshan, Chong-qing city's minority blood cell studying, this can be led to a new researching scope in the future anthropology, the scheme is a brand-new idea for anthropologists, etc.

The Two-dimensional generalized Korteweg-de Vries Burgers Equation With a forcing function is discussed, the existence and uniqueness of periodic traveling wave solution are proved.

Some results and new corollary to coincidence point theorems are obtained which extends the results in [1] and [2] ,Some new results are also given for Mann type Iteration with tow mappings in Banach space.

The K functional method is used in the paper which first describes the approximation and saturation problem by a certain trigonometric interpolation polynomials in the Besov spaces, the saturation class and the saturation order of approximation are determined.

In this paper, some linearization problems and the correlative symmetries of are discussed. By using Lie-Backclund transformation, some recursion operators of is got.

According the characteristic of radar video echo signals, the theory of wavelet multi-resolution decomposition and singularity detection is used in feature extraction of radar video echo signals, the amplitude and site of the modulus maximum point at different scales for radar video echo signals are obtained. These features and BP genetic algorithm are used to do the classifying identification, and the simulation results show that this method can get perfect effect.

Wavelet transform-based volume rendering has become one of the hot fields in research of visualization for the development of computer technology. On the basis of analyzing the volume rendering technique this paper reviews wavelet-based volume data representation and the application of wavelets to volume rendering, including hierarchical representation of volume data, wavelet-based compression and wavelet-based image order rendering as well as wavelet-based object order rendering, then evaluates their characteristics. Finally prospects research foreground of wavelet applications in volume rendering.

No-linear differential equations for hydraulic transient process, which is resulted from small increment of inlet pressure of pipe, have been conducted for turbulent flow. It is assumed that pressure and discharge after occurrence of the hydraulic transient process, are equal to an initial value added an additional pulse value respectively. The no-linear, unsteady partial differential equations for pressure and discharge have been converted into linear partial differential equations by the method of linearization treatment. Theoretical solutions are obtained by means of variable separation method. The initial, ultimate discharge and pressure at t = 0 and t → ∞ conforms to the reality.

In this paper, we give a class of explicit exact traveling wave solution to the two dimensional korteweg-devries burgers Equations

The efficient realization of Rijndael algorithm always is a hotpot in the research field. In this paper authors at first analyze encryption process of Rijndael, and then put forward speedy encryption algorithm based on checking-table. The most important means is constructing the list-table and skillfully using the s-table. Then we can encrypt plaintext by twice checking s-table or list-table and 12 times simple circle shift and sum modulo 2, thereby we can improve the realization efficiency of Rijndael algorithm. At the same time we analyze Rijndael algorithm's excellent performances in contrast to DES.

In this paper, we give some results for the blow-up behaviors of solution in finite time to the mixed problems for higher nonlinear hyperbolic evolution equations. By introducing the "blow-up factor K(u,u_t)" we get some new results it extensions the conclusions in [4,5] and [6].

Wavelet entropy (WE), a new complexity measure of electroencephalographic (EEG) signal, was investigated under two physiological states for a better understanding of brain dynamics. Id the quantitative analysis, higher mean WE values of 16-channel EEGs under eyes open state imply higher degrees of EEG complexity compared with their counterparts under eyes closed state, which exhibits that more complex process will occur even when a very simple neural information processing happens (e.g., eyes open). In the qualitative analysis, blocked alpha rhythms were observed under eyes open state over posterior (particularly occipital) areas via the time-frequency, the time-space, and the time-frequency-space variability of relative wavelet intensity (RWI); and the evolution of delta rhythm is against the change of the WE under both two states, which suggests stronger delta activities are corresponding with lower degrees of EEG complexity (i.e., lower WE).

The objective of this paper is to develop a family of one-dimensional wavelet-based finite elements. First, independent wavelet bases are used to approximate displacement functions, unknown coeficients are determined through imposing the continuity, linear independence, completeness, and essential boundary conditions. A family of one-dimensional wavelet-based shape functions are then developed, which are hierarchical due to multiresolution property of wavelet. Secondly, to construct one-dimensional wavelet-based finite elements, derivation of the shape functions for a subdomain is employed. Thus, the wavelet-based finite elements being presented are embodied with properties in adaptivity as well as locality. Numerical examples are used to illustrate the characteristics of the current elements and to assess their accuracy and efficiency.

In this paper, we present an approach to the detection and recognition of a building in natural environments based on its geometric features. We first carry out some preprocessing steps to the original images including binary-conversion, automatic horizontal segmentation, smoothing and edge detection. Then we use improved HOUGH transform and some line analysis strategies to do shape detection and recognition in order to get the potential regions of the target. Finally, we discard the false targets and find out the exact result. Experimental results show that this approach achieves excellent performance (not only makes no mistakes but also show a precise location and region of the target) to images when scaling, different angles of view, changes in conditions of luminance and the mosaic phenomenon are involved. The time and space complexity of it are feasible. Furthermore, this method is also suitable to other targets, which have regular and space invariant geometric features.

In this paper, we study a kind of Hopfield neural network, and present new sufficient conditions on Global Asymptotic Stability (GAS) and Global Exponential Stability (GES) by constructing the Liapunov function and utilizing the technique of inequality analysis, as well as improve upon the existing results.

Author Index.