This paper applies Parzen-window nonparametric estimation of active contour methods for image segmentation which can be performed without regarding image regions as any of typical probability distributions, so it expands the application range of the algorithm. In order to use the information of gray value congruence in a particular region, the Parzen-window parameter is alterable in estimating the densities. This will be more suit for practical image segmentation. Manj’ examples of synthetical and real image segmentation prove that the method is rationality, better for curve evolution and has good accuracy and antinoise function.

Firstly, order morphological transformations on gray image are introduced in this paper, then an image processing system in which order morphological transformation is implemented by hardware is presented. The system is composed of Digital Signal Pro-cessor (DSP) and Field Programmable Gate Array (FPGA), one FPGA as coprocessor can implement different functions due to its re-configurable characteristic. The images processed by software and hardware realization are compared in effect and speed. The algorithm is implemented successfully and its speed is very high. It makes possible that mathematical morphology is applicable in real time image processing.

Skin color character is widely used for face detection and face tracking. This paper construct a mixed model in both normalized RGB and HSV color space, which can be used to acquire possible skin regions in original image. It designs a sequential flexible morphology operators which can respectively eliminate noises and disturbances, separate face candidates, acquire eye and mouth "hole", and gain contours of faces. The experimental results show that the proposed approach can not only effectively eliminate human hands and arms non-face objects in image, but also smooth away difficulties of face occlusion, pose and orientation, size difference and light variety. Furthermore, this approach is easy to implement, and its computational spending is relatively smaller. So it is suitable to fast face detection.

The classical Principal Component Analysis (PCA) for image feature extraction is usually based on vectors, which makes it very time-consuming, and the class information in the training sample has not been utilized fully also. To overcome these two drawbacks of PCA, this paper proposes a novel and efficient PCA method based on original image matri-ces directly. It can extract the discriminant information included in the class mean images. Hence, the proposed method has better discriminant performance than classical PCA. Ex-perimental results on ORL face database show the proposed method is more powerful and efficient than the classical PCA and Fisher linear discriminant analysis.

In this paper a watermarking scheme for digital image that can resist geometrical attacks is proposed. The watermark is a key-dependent random vector according to Gaussian distribution. After preprocessed and periodized, the watermark is embedded into cover image in spatial domain exploiting the properties of HVS. The cover image is not needed in watermark detection process. First, compute the Autocorrelation Function(ACF) of estimated watermark. The peaks of ACF are used to determine the geometrical attacks applied to the stegoimage. Experimental results show the scheme is robust to compression, filtering, cropping as well as geometrical attacks.

Based on modification of time constraints in the FPGA design, a watermark-embedding scheme is proposed for protection of intellectual property rights of the system designer. A coded binary watermark sequence is used to replace the least significant digits of the time constraints in some non-critical paths. The modified time constraints lead to substantial and unique changes in the generated bit stream without altering the performance of the design, both in terms of space and time overheads and the system functionality. The embedded data can be extracted in a reverse procedure. The paper provides a scheme with zero area and low timing overheads, and a significant increase in embedding capacity in comparison with some existing techniques.

Based on relative error covariarice matrix (RECM) information, a reduced-order model is proposed for solving linear inverse problem. The reduced-order model turns the high order model into an approximate lower order model, which can efficiently alleviate the computational load of the inversion algorithm. Thus, the computational complexity difficulty arose in the solution of linear inverse problem can be conquered, and this in turn promotes the implementation of the inversion algorithm. In addition, the reduced-order model can improve the estimate precision of those points that provide significant information to the reconstruction of the object.

The real-time detection algorithm of moving point target in image sequences is one of key algorithms in many real-time processing systems. After the analysis of real background and the demand of the detection algorithm, this paper presents a research of the dynamic programming with LS linear predictor for detecting low SNR moving targets. With the result of simulation test, it can be pointed out that the introduced algorithm can effectively suppress the target energy pervasion and reduce the false alarms around the real target. The result is very suitable for the point-track predicting and matching algorithm for detecting the target trajectory.

After analyzing the inherence of radiometric resolution in SAR images, this paper points out there is considerable error of radiometric resolution in SAR images using current engineering estimation formula which only suits for the qualitative analysis. Based on probability of error in the theory of signal detection, an engineering estimation formula of radiometric resolution in SAR images is derived. The corrected formula suits for the quantitative analysis.

This paper amply analyzes the airborne SAR motion compensation mechanism on the range compression data and complex image data, and proposes a motion compen-sation algorithm for the high resolution airborne SAR. This algorithm makes use of the sub-aperture technique to estimate the sub-aperture’s Doppler rate with the improved max-imum contrast method, and then performs time-domain motion compensation for range compressing data and executes the frequency-domain motion compensation for complex im-age with PGA method. The raw data imaging result shows that the algorithm can effectively overcome the disadvantage of the scene dependence of the singularity motion compensation algorithm, and improve the SAR image quality.

The Wigner-Ville Distribution (WVD)-Hough transform could effectively detect several Linear Modulated Frequency(LMF) signals in low Signal to Noise Ratio (SNR.) envi-ronment. However, when the energy of these LMF signals diverge greatly, it is quite difficult to detect all the LMF signals with the WVD-Hough transform at the same time, because the platform of strong signals will cover weak signals. In this paper, a signal detection method based on Modified Smoothing Pseudo-WVD (MSPWVD) and binary integration in Hough transform parameter space is proposed according to the WVD-Hough transform. Using this method, strong signals as well as the weak signals covered by the strong signals can be de-tected simultaneously. It means that this method has high practicable value. Simulation results verify the effectiveness of this method.

Inspired by the biology immune system, the concept of Immunodominance in immunology is introduced to Artificial Immune System (AIS) in this paper. In AIS, the concept includes two parts, antibody-immunodominance and antigen- immunodominance. Based on the clonal selection theory, a new hybrid AIS algorithm, namely IrnmunoDominance Clone Algorithm (IDCA), with antibody-immunodominance is put forward. Compared with gene algorithms, IDCA is shown to be an evolutionary strategy capable of solving complox machine learning tasks, like 0-1 knapsack problem.

The principle of pseudo-random phase modulation fuze, pseudo-random phase modulation and Pulse Amplitude Modulation (PAM) combined fuze, and noise FM jamming are introduced in brief. On this basis, the Signal to Jamming Ratio (SJR) gains of the two fuzes in noise FM jamming environment are deduced in detail. Then, the working process of the two fuzes is simulated, whose results are close to the theory. These results and the output of correlation prove that the pseudo-random code fuzes have strong performances of anti-noise FM jamming generally.

It is a basic problem for the simulation of correlated Rayleigh. fading to produce the multiple independent Rayleigh fading waveforms. Jakes model has been widely accepted, but it has major shortcoming in the statistical independence. A new math model of multiple independent correlated Rayleigh fading is put forward. Both theory and simulation show that it is superior to the original Jakes model and its modified models that have been discovered in the aspects of the computation complexity and statistical independence.

A new selected transmission scheme of multiple antennas is proposed, named Greedy Search (GS) method. The transmit antennas selected from all antennas can be used as the active transmit antennas. When, used with waterfilling power allocation,’ GS method can decrease the complexity of signal design on antennas, compared with Waterfilling power allocation on all Antennas (WA). The capacity loss over WA in independent fading channels is rather small, moreover, this method can get capacity gain over WA in spatial dependent fading channels. The cost of capacity gain of GS is the necessity of Schmidt orthogonalization on the channel matrix. GS can be used when the channel matrix is not column-full-rank, i.e., the number of transmit antennas is greater than that of receive antennas, and the transmitter can get the full channel matrix information (used with waterfilling power allocation); or; the receiver can get the full channel matrix information (used with equal power allocation).

The paper studies the performance of CDMA systems with the space-fcime ap-proach under the conditions of multi-users and multipaths. The space-time spreading scheme with a novel algorithm can improve the downlink performance by using two antenna elements at the base station and one antenna at the mobile terminal. Different from the present other results, the multi-user systems have been analyzed, the detailed theory analysis and expression of bit-error-probability have been provided, based on the scheme of single-user systems. Moreover, the simulation results show that this scheme obtains an improvement of the CDMA system performance comparing to the conventional approaches, demonstrat-ing that it is a practical way to decrease the bit error rate and improve the qiiality of the downlink of CDMA systems. Meantime, the implementation complexity of the systems has increased, there are many new technical requirements for mobiles to be satisfied, and the STS technique needs to be improved.

The rate-distortion theory is induced to the system model, and the mathematic formula between the feedback channel rate and the closed-loop MIMO capacity are deduced. Through theoretical analysis and simulation results, this paper proposes four important conclusions on the relation between the feedback rate, the average SNR, the antenna number, and the closed-loop capacity. The conclusions, which are important to system design, provide the instructions to design and tradeoff the overhead of feedback channel and the capacity gain.

It is usually assumed that the perfect propagation delay has been achieved between the receiver and the transmitter, the receiver and the transmitter synchronize per-fectly, actually it is impossible to do this. Because of this, this paper proposes a single user TH-SS PPM UWB system with synchronization timing tracking ability, designs the synchronization timing tracking scheme, derives the receiver demodulator output SNR while the synchronization timing error exists between the transmitter and the receiver. The syn-chronization timing error has great effects on the system BER performance. Theoretical analysis and simulation results indicate that the less the synchronization timing error, the less the system BER.

Ad hoc network is an energy-limited network. How to improve energy efficiency in an Ad hoc network is an important open problem. This paper proposes a new routing metric, which integrates congestion status at the link layer and remaining energy of a node. By adopting this new metric, energy dissipation and traffic load are distributed more evenly in the network, leading to higher network connectivity and network throughput. Simulation results show that the new metric improves network throughput for networks with different scales.

In this paper, a transmission time based admission control algorithm in IEEE 802.11 wireless LANs is proposed. When a new real-time flow requests to be admitted to the system, the Access Point (AP) evaluates whether the new flow will overrun the system transmission time budget. For the admitted flow, the AP adopts a token bucket to allocate the transmission time and polls the flow regularly, so the quality of service of the flow will be guaranteed. Furthermore, by means of excluding extra real-time flows, some wireless resources will be left for non-real-time services. The fairness of wireless resources sharing will be satisfied.

Firstly, the uplink capacity of CDMA system, in multi-path fading channels with single-user-detection and multi-user detection, is analyzed in this paper. The uplink capacity comparison between LS code and Walsh code is discussed for the above-mentioned receivers (multi-user detector, single-user detector) in multi-path fading channels. The result shows the capacities of LS code and Walsh code are approximately the same when the multi-user detector is adopted. However, the capacity of LS code is much higher than Walsh code in condition that single-user detector is applied.

A novel model of the multiplex system is proposed, which can suppress narrowband interference and white Gaussian noise by using adaptive combination filter and estimation model of SPECCORR_TDOA based on analyses of cyclostationary signals. Finally, the multiplex system models are simulated with computer. The result proves that the improved multiplex system has strong suppression ability of interference and noise.

Because the algebraic expression of Rijndael S box is a composition of the converse function with a q-polynomial over GF(2⁸), in this paper the variables of S box are supposed rationally and the relations between these variables are analyzed, then a new system of multivariate quadratic equations over GF(2⁸) are used to describe completely Rijndael, the cryptanalysis of Rijndael can be written as a problem of solving the system of multivariate quadratic equations. This system is simpler than Murphy and Robshaw’s, and has a lower complexity while applying XSL technique.

Analyses of a nominative signature scheme (KPW scheme) and a nominative proxy signature scheme (PL scheme) are given, respectively. The results show that the KPW scheme is not a nominative signature scheme since the nominator not only can verify and prove the validity of given signatures but also can convert given signatures into universally verifiable signatures, and that the PL scheme is only a self-authentication signature scheme since anyone can verify the validity of given signatures. Furthermore, an improvement of the KPW scheme, in which all properties of the nominative signature are held, and a nominative proxy signature scheme based on the proposed nominative signature scheme are proposed.

The reflector antenna distortion model is built by exploding test results. The radiation characteristics of a kind of distorted shaped reflector antenna are computed by using PO and PTD method, and the radiation characteristics of the reflector antenna with holes on the surface are also computed. The simulation results show that because of the local distortion of the reflector the radiation patterns become degraded. The sidelobe level in the azimuth plane becomes higher as the distortion area becomes bigger. It can be seen that the sidelobe level in the azimuth plane also increases because of the reflector perforated with holes on the surface.

This paper surveys some typical approaches to migration models used in Mobile Agent System(MAS) and describes the migration mechanism in MAS which is designed and implemented by authors. Focused on the issues about the agent’s dynamic routing and migration-related security changing, an extended migration mechanism of agent is proposed and an application example is given.