In order to solve the problem of human aggressive behavior recognition, an aggressive behavior recognition method based on human joint points was proposed. Firstly, OpenPose was used to obtain the human joint point data of a single frame image, and nearest neighbor frame feature weighting method and piecewise polynomial regression were used to realize the completion of missing values caused by body self-occlusion and environmental factors. Then, the dynamic "safe distance" threshold was defined for each human body. If the true distance between the two people was less than the threshold, the behavior feature vector was constructed, including the human barycenter displacement between frames, the angular velocity of human joint rotation and the minimum attack distance during interaction. Finally, the improved LightGBM (Light Gradient Boosting Machine) algorithm, namely w-LightGBM (weight LightGBM), was used to realize the classification and recognition of aggressive behaviors. The public dataset UT-interaction was used to verify the proposed method, and the accuracy reached 95.45%. The results show that this method can effectively identify the aggressive behaviors from various angles.

In the non-rigid point set registration algorithm, whether the correct correspondence can be found plays an important role. Generally the corresponding points in two point sets have similar neighborhood structures besides the close distance. Therefore, a non-rigid point set registration algorithm based on global and local similarity measurement was proposed. Firstly, the Coherent Point Drift (CPD) algorithm was used as the registration framework, and the Gaussian mixture model was used to model the point sets. Secondly, the global and local mixture distance was improved to form the global and local similarity measurement criterion. Finally, the correspondence and the transformation formula were solved by the Expectation Maximization (EM) algorithm. In the initial stage of the iteration, the proportion of local similarity was larger so that the correct correspondence was able to be found rapidly; with the progress of the iteration, the proportion of global similarity was increased to ensure the smaller registration error. Experimental results show that compared with the Thin Plate Spline Robust Point Matching (TPS-RPM) algorithm, the Gaussian Mixture Models point set REGistration (GMMREG) algorithm, the Robust Point Matching algorithm based on L ₂E estimation (RPM-L2E), the Global and Local Mixture Distance and Thin Plate Spline based point set registration algorithm (GLMDTPS) and the CPD algorithm, the proposed algorithm has the Root Mean Squared Error (RMSE) decreased by 39.93%, 42.45%, 32.51%, 22.36% and 11.76% respectively, indicating the proposed algorithm has better registration performance.

To address the problems that data transmission delay is high and the channel utilization rate is low in the TAB-MAC (Assisted Beamforming MAC (Medium Access Control) protocol for terahertz communication network) for present terahertz Wireless Personal Area Network (T-WPAN), an Efficient and Fast dual-channel MAC protocol for T-WPAN (EF-MAC) was proposed. A test frame was sent to the source node through destination node to reduce an acknowledgment frame, so as to reduce control overhead and test delay. And then a sending and receiving mechanism of node location information was adaptively concelled, the source or destination node obtained the location information of the other node through interaction process previously of RTS/CTS (Request To Send/Clear To Send) frame, and the position of other node had not changed, thus the location information of RTS or CTS frame could be omitted to reduce control overhead. The theoretical analysis and simulation results show that compared with TAB-MAC protocol, the proposed protocol can effectively reduce data transmission delay and improve network throughput.

In cognitive satellite terrestrial networks, when the satellite users are secondary users, power control is necessary to guarantee the communication quality of terrestrial primary user in the uplink case. In the context of fading channels, maximizing the Ergodic Capacity (EC) of the satellite user was selected as the objective function, then two optimal power control schemes were proposed based on Peak Interference power Constraint (PIC) and Average Interference power Constraint (AIC), respectively. Meanwhile, the closed expression of the optimal transmit power was given. Simulation results show that the ergodic capacity of satellite user can be increased when the satellite link experiences the weaker shadowing conditions; moreover, under the the specific satellite link condition, the performance of satellite user becomes better with the increasing of the terrestrial interference link fading parameters. In addition, power control method based on AIC is superior to that based on PIC.

Because of the successful use of the sparse representation in face classification algorithm, a more efficient classification method based on Sparse Representation-based pattern Classification (SRC) and elastic network was proposed. To enhance the ability of collaborative representation and enhance the ability to deal with strongly correlated data, a sparse decomposition method based on elastic network was proposed based on the iterative dynamic culling mechanism. Test samples were represented by a linear combination of training samples, and the iterative mechanism was used to remove the categories and samples with less contribution to the classification from all the samples, the Elastic Net algorithm was used for coefficient decomposition to select the samples and classes with high contribution to the classification. Finally, the test samples were classified according to the similarity. The experiment results show that the recognition rate of the algorithm is 98.75%, 86.62% and 99.72% respectively for the ORL, FERET and AR data sets which shows the effectiveness of the proposed algorithm. Compared with the methods of LASSO and SRC-GS, the proposed algorithm can enhance the ability of dealing with high-dimension small sample and strongly correlated variable data in the process of coefficient decomposition. It highlights the importance of sparse constraint in the algorithm and has higher accuracy and stability, and can be more effectively applied to face classification.

Aiming at the problem of object structure discontinuity and incompleteness occurred in image completion, an image completion algorithm based on depth information was proposed. Firstly, the plane parameter Markov random field model was established to speculate depth information of the pixels in the image where the scene situate, then the coplanar region in the image determined, and the target matching blocks were located. Secondly, according to the principle of perspective projection, the transformation matrix was derived, which guided the geometric transformation of the matching blocks. Finally, the target cost function which includes the depth term was designed. Experimental results show the proposed algorithm has superiority in both subjective details and Peak Signal-to-Noise Ratio (PNSR) statistics.

The traditional patch-based image completion algorithms circularly search the most similar patches in the whole image, and are easily affected by confidence factor in the process of structure propagation. As a result, these algorithms have poor efficiency and need a lot of time for the big computation. To overcome these shortages, a fast image completion algorithm based on randomized correspondence was proposed. It adopted a randomized correspondence algorithm to search the sample regions, which have similar structure and texture with the target region, so as to reduce the search space. Meanwhile, the method of computing filling priorities based on confidence factor and edge information was optimized to enhance the correctness of structure propagation. In addition, the method of calculating the most similar patches was improved. The experimental results show that, compared with the traditional algorithms, the proposed approach can obtain 5-10 times speed-up in repair rate, and performs better in image completion.

Aiming to the practical problem that it is difficult to estimate the Richards model parameters, the parameter estimation problem of the Richards model was formulated as a multi-dimensional unconstrained function optimization problem. Combined with the actual growth concentration of glutamic acid, in Matlab 2012b environment, the fitness function was established by Particle Swarm Optimization (PSO) algorithm, four parameters of Richards model were estimated by the least square method, and the growth curve and the optimum curve were established. To further verify the effectiveness of the algorithm, the PSO algorithm was compared with traditional parameter estimation method, such as four point method and Genetic Algorithm (GA) method, the related index and the residual standard deviation were used as the evaluation index. The results show that, the PSO algorithm has better fitting effect for Richards model and good applicability for parameter estimation.

With reference to the properties of multiscale and shift invariance of nonsubsampled Contourlet transform, and concerning the characteristics of Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) images, a medical images fusion method was proposed．The proposed method fused the low frequency subband and high frequency subband of these medical images separately by the regional feature strategy. The paper introduced the judgment criteria of images fusion and expatiated on the principle and implementation of Nonsubsampled Contourlet Transform (NSCT). And this gave the subjective judgment and numeric measurement of the fusion images based on visual effect and information indexes. To evaluate the performance of the proposed algorithm, the authors compared the results with those of the algorithms, such as wavelet transform and Contourlet transform. The CT and MRI images simulation results of mandibular system indicate that the proposed method outperforms the others in terms of both visual quality and objective evaluation criteria, while it can integrate and maintain much more effective and detailed information as well.

A free search algorithm was proposed for the discrete optimization problem. However，solutions simply got from free search algorithm often have crossover phenomenon. Then, an algorithm free search algorithm combined with cross elimination was put forward, which not only greatly improved the convergence rate of the search process but also enhanced the quality of the results. The experimental results using Traveling Saleman Problem (TSP) standard data show that the performance of the proposed algorithm increases by about 1.6% than that of the genetic algorithm.

The research on the evolutionary mechanisms and models of Protein-Protein Interaction (PPI) network is significant for understanding the evolution of the biological systems as well as the formation process of the organisms. So far, there have been kinds of models based on different evolutionary mechanisms. All of these models exhibit certain topological characteristics emerging from the protein-protein interaction networks, while some limitations exist simultaneously. This paper focused on several classic protein-protein interaction network models, analyzing the main ideas of these models and comparing the topological characteristics derived from them with those of real protein-protein interaction networks. A summary of the features for each model was given based on the experiments. At last, several viewpoints for the future research of protein-protein interaction network models were also proposed to provide a useful reference for further studies.

Concerning the faults of local extremum in image registration based on mutual information, a new medical image registration method based on Powell and improved genetic algorithm was proposed in this paper. It put forward an improved method regarding the shortcomings of the standard genetic algorithm, such as slow convergence and prematurity that will result in artifacts, and generated the iteration individual by Logistic chaos map. This method utilized the multi-resolution analysis strategy and searched for the optimal of the objective function by this hybrid optimized algorithm in the lowest resolution image level. Then it continued the optimization course and accomplished the image registration by this optimal data with the Powell algorithm. The experimental results indicate that this algorithm can effectively improve the image registration velocity and avoid local extremum of the operator while getting better performance of image precision in contrast to the Powell algorithm and unimproved genetic algorithm.

It is a difficult and hot topic in video surveillance to detect moving objects with moving camera．In order to detect moving objects effectively,according to the characteristics of the different speed between the background and moving target, a method was proposed based on Scale Invariant Feature Transform （SIFT） features matching and K-means clustering．The SIFT features of the two adjacent frames in the video were extracted and matched firstly. After that the velocity of the matched SIFT features were computed. Finally the K-means clustering method was used to analyze the SIFT features of the moving objects and background and experiments were done in the cases of single moving object and multiple moving objects and when the camera was rotated. The experimental results demonstrate that, the proposed method can detect targets effectively and remain the stable local features of targets in moving background and have good adaptability to changing illumination and camera movement and rotation．

In order to reduce the number of triangular patches in surface rendering, save space and improve reconstruction results, a contour connecting method based on interzone line was proposed in this paper. This connecting method needed to resample with angle and generate secondary resampling points with geometric methods. And resampling points corresponded with each other. Tissue surfaces were reconstructed with quadrilateral patches and 3D reconstruction was realized. In experimental simulation, a series of head CT images were researched. The simulation analysis shows that to resample the contour line and use quadrilateral patches instead of triangular ones, the number of patches is reduced and the storage space is saved. The reconstruction time is short and the reconstruction efficiency is high.

On clustering large data, the spectral clustering can hardly be promoted because of large occupied storage space and high time complexity. Hence, multi-segment and upward and downward double-direction shrink QR algorithm for the corresponding eigenvectors of eigenvalues was adopted to achieve dimensionality reduction. Then, a new quantum genetic spectral clustering algorithm was proposed to cluster the sample points in the mapping space. Compact input with low-dimension for quantum genetic spectral clustering was obtained after mapping, and the quantum genetic spectral clustering algorithm, characterized by its rapid convergence to global optimum and minimal sensitivity to initialization, can obtain good clustering results. The experimental results show that the proposed method is superior to the spectral clustering algorithm, K-means, NJW algorithm in astringency and stability and has a higher overall optimal solution.

A new condition prediction approach based on weighed Hidden Markov Model (HMM) was presented in order to solve the problem of trend prediction for vehicle power system. Through the establishment of the power system HMM, the hidden state of HMM was predicted by weighted prediction method, and the observed state of the model was calculated by the observation probability density of the maximum probability hidden state. The approach was applied to the state prediction of rate turn of the system voltage adjusting pulse width signals, and was compared to that of using BP (Back Propagation) neural network and Auto-Regression (AR) prediction model with the same sequence. The results show that the method has better prediction on the system state change.

To solve the problem of transmitting credentials and other resources through unsafe physical channels during an Automated Trust Negotiation (ATN), a transmission scheme for credentials and resources was proposed based on magic cube algorithm. Through the magic cube algorithm, a transformation sequence was formed in terms of the request or the resource of negotiation initiator, followed by the digital digest to generate the information transformation sequence. According to the logical expression composed of credentials which represent the condition negotiation success, the information transformation sequence was shuffled to form an information transmission sequence, which was sent to the negotiation receiver. The information transmission sequence was reciprocally transformed by the negotiation receiver according to his own credentials. This scheme has many features of the one-round credential exchange, and little network cost. The example shows that the scheme is feasible, and the experimental results show that the scheme has good security and efficiency and low information transmission capacity.

The accuracy of sensor data is a critical index to evaluate the performance of Wireless Sensor Network (WSN). Outlier detection is a crucial but challenging issue for WSN. In this paper, an outlier detection approach based on variable-width histogram was proposed. The dynamic sensor data were aggregated into variable-width histograms, which avoided unnecessary data transmissions while detecting outliers. The theoretical analysis and evaluation on real WSN dataset show that this approach has high detection accuracy, and the cost is effectively reduced.

In order to modify the incomplete road extraction caused by shadow, shelter and noise in high resolution remote sensing images, a Ribbon Snake model with width information was established based on the geometric characteristics of road. To overcome the great dependence of interior parameters and the easy being affected by the complex background of Ribbon Snake, a B-spline Ribbon Snake was constructed, where the smoothness of the Snake was implicit in the B-spline formulation and the flexibility of the Snake was adjusted by the number of control points. The road network segmentation results show that the improved B-spline Ribbon Snake can obtain a more accurate and smoother segmentation and is more robust to noise.

DDoS attack is hard to detect in backbone network, for the reason that attack flows are distributed in multiple links and prone to be masked by tremendous amounts of background traffic. To solve this problem, a detection method based on global abnormal correlation analysis was proposed. The change of correlation between traffic caused by attack flows was exploited for attack detection, the correlation between potentially anomalous traffic was extracted by principle component analysis, and its change degree was used as an indicator of attack. Evaluation shows effectiveness of the proposed method, and proves that it overcomes the difficulties in detecting relatively low volume of DDoS attack transiting in backbone network. Compared with the existing network-wide detection method, it achieves higher detection rate.

The chip of smart card stores private sensitive defense data of user, which concerns the benefits of customers and mobile phone service providers. To enhance the security and correctness of data with a limited capacity chip is a critical problem. The security module architecture of chip operating system in evolution data of CDMA2000 was designed, and the network access authentication algorithm was put forward with a space optimization method. The file access control, message authentication and data encryption for communication were designed, and the power down protection while modifying flash memory or update several files was introdued. Experimental results prove that the system can run stably, effectively and safely, and ensures the integrity, validity, authenticity of data in storage and transmission between smart card and network.