Search

For the definition of heterogeneous and non-isotropic medium diffusion equation, we have established a finite-difference algorithm for the numerical solution of diffusion problems. This algorithm not only applies to the convex meshes, but also applies to the concave grid or self-AC grid. The performance of this algorithm is comparable to other algorithms for problems with smooth coefficients and regular grids, but it is superior to the problems with rough coefficients and no-skewed grids. We will discuss the truncation error and correction algorithm of the support-operator in this article.

First in this paper, we will introduce a map matching algorithm based on fuzzy logic, and how we use it to fuse the navigation data. Then we will introduce a kind of iterative algorithm what we use to optimize the parameters of the map matching algorithm we mentioned before. So we can improve the precision of our map matching algorithm.

Social network is a collection of heterogeneous multi-relational data represented by the graph, whose nodes represent object, whose edges represent relationships between nodes, and the weights represent the extent of the relationship between nodes. This paper gave a weighted K-means algorithm and introduced weighted K-means algorithm into social networks. Traditional k-means and most k-means variants are still computationally expensive for large datasets, however, the weighted K-means algorithm is to reduce the initial cluster centers blindness and randomness by eliminating noise point and narrowing the range of k values. Experiments datasets show that the weighted K-means algorithm significantly enhances the clustering quality. Therefore, the weighted K-means algorithm is effective and suitable for the social network. Algorithm’s error rate is smaller and accuracy is higher than that of traditional k-means algorithm.

The localization of anchor-free wireless sensor network’s serious cumulative error leads to low positioning accuracy. To solve this problem, this paper proposed a co-located positioning method based on asynchronous change of learning factor adaptive weights particle swarm optimization algorithm (SAAPSO) and Taylor algorithm. The first phase of localization is building relative coordinate system. The second phase is estimating the node’s initial position with SAAPSO algorithm and using Taylor algorithm to iterative calculate in the node’s initial position to obtain accurate result. In the third stage, the node with precise coordinates takes part in the localization of rest nodes. Simulation results show that: this positioning algorithm has smaller cumulative error and higher accuracy.

Image matching plays an important role in computer vision. The features extracted by SIFT algorithm have high stability invariant to scale, rotation and light, so it is the most popular algorithm for image matching. While SIFT algorithm also has its disadvantages of high dimensional data and time-consuming. To solve this problem, the traditional method employs PCA algorithm to reduce dimensionality of the descriptors. While PCA is a linear dimensionality reduction algorithm which means that it can only be used for linear distributed data. This paper employs the fuzzy K-means algorithm to improve it (referred to as FKPCA) and improved RANSAC algorithm to eliminate false matching points after matching with PCA-SIFT and FKPCA-SIFT. From the experimental results, compared with PCA-SIFT algorithm, it can be seen that FKPCA-SIFT can keep the high matching accuracy for image. Moreover, FKPCA-SIFT can also be applied to non-linear data to expand the scope of PCA-SIFT and provides a better reference platform for further research.

The problem of local minimum cannot be avoided when it comes to nonlinear optimization in the learning algorithm of neural network parameters, and the larger the optimization space is, the more obvious the problem becomes. This paper proposes a new type of hybrid learning algorithm for three-layered feed-forward neural networks. This algorithm is based on three-layered feed-forward neural networks with output layer function, namely linear function, combining a quasi Newton algorithm with adaptive decoupled step and momentum (QNADSM) and iterative least square method to export. Simulation proves that this hybrid algorithm has strong self-adaptive capability, small calculation amount and fast convergence speed. It is an effective engineer practical algorithm.

Ranking data points with respect to a given preference criterion is an example of a preference learning task. In this paper, we investigate the generalization performance of the regularized ranking algorithm associated with least square ranking loss in a reproducing kernel Hilbert space, and use the method of computing hold-out estimates for the proposed algorithm. Based on using the hold-out method, we obtain fast learning rate for this algorithm.

Let s and t be two points on the boundary of a simple polygon, how to compute the Euclidean shortest path between s and t which visits a sequence of segments given in the simple polygon is the problem to be discussed, especially, the situation of the adjacent segments intersect is the focus of our study. In this paper, we first analyze the degeneration applying rubber-band algorithm to solve the problem. Then based on rubber-band algorithm, we present an improved algorithm which can solve the degeneration by the method of crossing over two segments to deal with intersection and in our algorithm the adjacent segments order can be changed when they intersect. Particularly, we have implemented the algorithm and have applied a large of test data to test it. The experiments demonstrate that our algorithm is correct and efficient, and it has the same time complexity as the rubber-band algorithm.

The diamond search (DS) algorithm is one of the most efficient block matching motion estimation algorithms by far and has already been applied in MPEG2/4. Through our research, we found that there is still some redundancy in the algorithm. In this paper, an improved new difference based search (DBS) algorithm is proposed. Simulation results demonstrate that the new algorithm outperforms the well-known diamond search (DS) algorithm and four step-searches (4SS). It obtains almost the same Peak Signal to Noise Ratio (PSNR) while requires less computations than the DS algorithm and 4SS algorithm.

K-Means algorithm is a one of the most famous unsupervised clustering algorithm. It has many disadvantages, such as sensitivity to the initial clustering centers and computes all the data points multiple times when facing the increasing data volume. In order to overcome the above limitations, this paper proposes to make use of density idea to find k cluster centers by adjusting the threshold. Finally, we design and implementation of the K-Means algorithm on the modern Graphic Processing Unit (GPU). The ratio of distance between classes to distance within classes and speedup are used as evaluation criteria. The experiments indicate that the proposed algorithm significantly improves the stability and efficiency of K-Means algorithm.

Fuzzy c-means (FCM) algorithm is an unsupervised clustering algorithm for image segmentation, and has been widely applied because the segmentation results are consistent with human visual characteristics. Enhanced fuzzy c-means clustering (EnFCM) algorithm is the improved FCM algorithm, which reduces the computational complexity. But, both FCM algorithm and EnFCM algorithm, clustering number still need to be manually determined. This paper, in order to realize the automation degree of algorithm, presents an improved algorithm. It first analyzes the histogram, then automatically determines the clustering number and peak value of each class through use of the peak point detection technology, finally segments image by using EnFCM algorithm. Experiments show that this method is a kind of faster fuzzy clustering algorithm with automatic classification ability for image segmentation.

A new Evolutionary algorithm was presented based on intermediate value theorem of continuous function. The global search ability and the local search ability of this algorithm are well balanced, and operators used in the algorithm are simple, Further, small size of population scale is used. Initial numerical experiments show that the mean evolutionary algorithm is better than differential evolution algorithm in solving high dimension function optimization.

Data simplification is an important factor of the spatial data generalization. In this paper, the principle of the 3D D-P algorithm was introduced firstly, and aiming at the low computational efficiency of this algorithm, the first base plane was artificially designated instead of the loop search in the tight method, this improvement makes the computational efficiency to improve dozens of times , but the result of generalization of 3D discrete point is very similar to the generalized result by the tight method. This process was validated by experiments and this algorithm has a certain overall characteristics,therefore it has wide application prospect.

The RL(Richardson-Lucy) algorithm is an important method for restoration of turbulence-degraded images. However, the shortcoming of this method is that it tends to amplify the noise and exsits excessive smoothing in the iterative procedure. This paper discusses the RL algorithm and its improving methods focusing on turbulence-degraded images restoration.Firstly, a short exposure atmospheric turbulence-degraded model is established and a numerical computing method is proposed for random phase screen. Secondly, the essential principle and computational formula are deduced. To restore the object image effectively from the turbulence-degraded image, a new double-circulation iterative Richardson-Lucy restoration algorithm using TV-regularized method is proposed. This new algorithm introduces the total variation restraint and estimates the object image and the point spread function based on the inner and outer double-circulation iteration, which can use the inherent relation between the object image and the point spread function adequately. Simulation experiments show that the proposed algorithm can effectively preserve the details and edges of the image and its restoration effect is obviously better than the traditional RL algorithm.

This paper presents a novel decoding algorithm for bit-interleaved coded modulation iterative decoding (BICM-ID) embedded turbo codes. It can yield good bit error rate (BER) performance with much lower complexity. The improved algorithm exploits a linear interpolation and optimal mean square approximation function to replace the logarithmic correction in the Jacobian logarithmic function based on the MacLaurin Series, which avoids complicated logarithm look-up table operations in Log-MAP. Simulation results show that the novel algorithm obtains can offer almost equivalent performance to the optimal algorithm. Compared with the improved MAX-Log-MAP algorithm, the proposed algorithm can reduce about 34% of computational complexity, meanwhile it achieves 0.1db-0.16db performance gains.

This paper studied a method which can extend Mean-Shift algorithm based on Kernel. We describe its basic idea, given the specific steps of the algorithm. And given the application of extended Mean-Shift algorithm in the area of image smoothing. The research also show that extended Mean-Shift algorithm has displayed strong vitality in image segmentation and real-time object tracking.

In this paper, two shortcomings of standard ART2/ART2A algorithm were revealed through theoretical analysis: (1)Standard ART2/ART2A algorithm is only suitable for the case in the nonnegative real number field because of a limit of pretreating process in F1layer; (2)Even through all input patterns are shifted to the nonnegative real number field through coordinate transformation, the standard ART2/ART2A algorithm can not correctly recognize those patterns which have same phase, but different amplitudes. As a result, the standard ART2/ART2A algorithm is not quite suitable for universal pattern recognition. So this paper presented a new nonlinear transforming function in F1layer and a new competitive learning formula in F2layer for traditional ART2/ART2A algorithm. The applicable scope of the new ART2/ART2A algorithm is expanded to entire real number field from nonnegative real number field. The result of typical calculation example shows that the presented algorithm is effective.

The number of fuzzy rules directly determines the complexity and efficiency of Fuzzy Neural Network (FNN). The Iterative Pruning (IP) algorithm belongs to the Pruning Method, and it spends much time computing adjusting factors of the remaining weights. So the Improved Iterative Pruning (IIP) algorithm is put forward, which adopts dividing blocks strategy and uses the Generalized Inverse Matrix (GIM) algorithm to replace the Conjugate Gradient Precondition Normal Equation (CGPCNE) algorithm for updating the remaining weights. The IIP algorithm is applied in the rule-reasoning layer of FNN to simplify its rules and structure in a great extent and preserve a good level of accuracy and generalization ability without retraining after pruning. The simulation results demonstrate the effectiveness and the feasibility of the algorithm.

Traditional K-means clustering methods have great attachment to the selection of the initial value and easily get into the local extreme value. This paper proposes a synthetic clustering algorithm of rough set and K-means based on Ant colony algorithm. While the rough set theory presents processing method of uncertain boundary objects, Ant colony algorithm is a bionic optimization algorithm, which has strong robustness, easily with other method unifies, solving efficiency higher characteristic.. Therefore, the K-means algorithm based on Ant colony algorithm in this paper combines rough set theory with simulated annealing algorithm and K-means, in which K means cluster number and initial cluster centers can be obtained dynamically with the principle of maximum minimum, and processing boundary objects with upper and lower approximation of rough set theory. Finally, the UCIs Iris set is used to test the algorithm. The experimental results show that the algorithm has higher accuracy rate, faster execution time and more stable performance.

An algorithm based on free search is proposed for the combinatorial optimization problems. In this algorithm, a feasible solution is converted into a full permutation of all the elements and a transformation of one solution into another solution can be interpreted the transformation of one permutation into another permutation. Then, the algorithm is combined with intersection elimination. The discrete free search algorithm greatly improves the convergence rate of the search process and enhances the quality of the results. The experiment results on TSP standard data show that the performance of the proposed algorithm is increased by about 2.7% than that of the genetic algorithm.

Fuzzy C-means (FCM) clustering algorithm is one of the widely applied algorithms in non-supervision of pattern recognition. However, FCM algorithm in the iterative process requires a lot of calculations, especially when feature vectors has high-dimensional, Use clustering algorithm to sub-heap, not only inefficient, but also may lead to "the curse of dimensionality." For the problem, This paper analyzes the fuzzy C-means clustering algorithm in high dimensional feature of the process, the problem of cluster center is an np-hard problem, In order to improve the effectiveness and Real-time of fuzzy C-means clustering algorithm in high dimensional feature analysis, Combination of landmark isometric (L-ISOMAP) algorithm, Proposed improved algorithm FCM-LI. Preliminary analysis of the samples, Use clustering results and the correlation of sample data, using landmark isometric (L-ISOMAP) algorithm to reduce the dimension, further analysis on the basis, obtained the final results. Finally, experimental results show that the effectiveness and Real-time of FCM-LI algorithm in high dimensional feature analysis.

In this paper, we study how to improve the efficiency of classic AC matching algorithm for scanning systems, and proposed an algorithm named AC-SUN based on Sunday matching algorithm. In our design, AC-SUN combines jumps strengths of Sunday algorithm and AC algorithm to avoid verbatim matching problem, meanwhile it retaining the advantages of the AC algorithm that establishes state tree and invalid pointer to increasing successful matching percentage. Our experimental result also verified its superiority to the classic AC algorithm.

In this paper, based on the combination of Genetic algorithm and BP algorithm, a new algorithm is proposed in this paper. The BP operator is embedded in the genetic operation in the algorithm, the algorithm effectively assimilates the global optimization of genetic algorithm and fast convergence of BP algorithm, and it encodes the construction and the weights hybrid with real code and binary code, achieving the same step optimization of structure and weights. The simulation results show that, the new algorithm can quickly converge to the global optimal solution, but also can obtain the best approximation of weights in the network structure.

Aiming at the problem of K-Means algorithm which is sensitive to select initial clustering center, this paper proposes a kind of initial point of K-Means algorithm. The algorithm processes the properties of the data objects, which determines the density of data object by counting the number of similar data objects and selects the center of categories according to the density of data object. The cluster numbers given and the UCI standard sets of data and the random data sets used, the clustering results demonstrate that the proposed algorithm has good stability, accuracy.

In this paper we combine the largest minimum distance algorithm and the traditional K-Means algorithm to propose an improved K-Means clustering algorithm. This improved algorithm can make up the shortcomings for the traditional K-Means algorithm to determine the initial focal point. The improved K-Means algorithm effectively solved two disadvantages of the traditional algorithm, the first one is greater dependence to choice the initial focal point, and another one is easy to be trapped in local minimum [1][2].

This paper presents a fault diagnosis method of BP neural network based on Levenberg-Marquardt learning algorithm. First, the use of principal component analysis to reduce the dimension of the fault sample reduced BP neural network input variables. Then use the Levenberg-Marquardt learning algorithm to adjust the network weights. Levenberg-Marquardt learning algorithm is combination of the Gauss - Newton algorithm and steepest descent algorithm. It has Gauss - Newton algorithm of local convergence and gradient descent algorithm of the global characteristic. So it has higher convergence speed, reduces the training time, to a certain extent, overcomes the problem of traditional BP network convergence speed slow and easy to fall into local minimum point. Simulation results demonstrate the correctness and accuracy of this fault diagnosis method.

For the purpose of improving the real-time performance of the SIFT algorithm, an image matching algorithm based on SUSAN-SIFT algorithm is proposed in this paper. Use SUSAN algorithm in the detecting feature points part of the SIFT algorithm, avoiding the time-consuming down-sampling and Gaussian convolution of the algorithm, and remove the feature points of the unstable and low-contrast by the methods of interpolated estimate and the principal curvatures, and then use the SIFT algorithm to achieve the parts of describing the feature points and images matching. experimental verification: the SUSAN-SIFT algorithm has more fast calculation speed than SIFT algorithm ensuring the accuracy at the same time.

In order to overcome the low efficiency of the database-based algorithm and the string-matching based algorithm in searching phone number, the two-level search algorithm by combining a decimal tree with a local registering table is proposed in this paper. The algorithm implementation is introduced in detail, and the experiments show that the decimal tree based algorithm is faster than the database-based algorithm and the string-matching based algorithm, especially much faster in the case of huge traffic.

This paper was enlightened by Douglas-Peucker Algorithm, using the algorithm in the application of railway vertical section automatic design. Firstly, the paper introduced the basic principle of Douglas-Peucker algorithm, based on which, the flow path of railway vertical section automatic design was formulated according to the design criterion. Secondly, the algorithm result was applied in Mengxi Coal Transportation Corridor, the quantity of the railway was acquired after automatic design with improved Douglas algorithm, through the quantity, and it could be proved that the algorithm did design a smooth slope. Automatic design in railway and road has become an inexorable trend, since it reduced the human cost, improved efficiency and avoided artificial defects.

An adaptive genetic algorithm was proposed to optimization bound in order to speed up the convergence of Gaussian mean shift. In practical question, as it's difficult to give a critical value definitely. According to the engine nonlinear of the corresponding oil and performance parameters, in gradient genetic algorithm, BP algorithm of local search is introduced. The adaptive value of chromosomes group gets quickly improved with the search in one coding field getting avoided due to the utilization of knowledge of chromosomes in problem-domain. The crossover and mutation operations are added so that chromosomes will not fall into the local minimum point in neighborhood. The experimental results prove that the convergence speed of the proposed method is non-linear and the use of gradient genetic algorithm is a fast algorithm that can support the global optimization of chromosomes in a group of process of iteration.

Through the arduous study of the famous Cohen-Sutherland algorithm, this paper proposes a novel improved algorithm by analyzing its main flaw in efficiency. The line segment without the clipping window will be rejected in our algorithm. By means of adding auxiliary straight-line, this algorithm efficiently calculates intersection points between line and the boundary of clipping window. Thus, this improved algorithm is simple, improving its clipping efficiency and reaching the desired results.

Because less data can be hidden on the small vector data set , its copyright is very difficult to obtain protection, and the data is vulnerable to attack, so a new double watermarking algorithm was proposed in this paper. Its main features are: 1) it selected feature points from the sequence of Douglas, then embedded watermarking points on both sides of the feature point, and last embedded watermarking points by wavelet transform again; 2) Further more it increased the map graphic deformation control design. The algorithm was applied to experimental data, and the test results showed that the algorithm had good robustness on graphics, rotation, scaling and shifting geometric transformation of points in graph layers. However, due to the smaller vector data set, so the algorithm is robustness on deletion and cropping of points in graph layers.

In image matching research, how to ensure that best match’s accuracy of the premise and a significant reduction in the amount of computing is the focus of concern by researchers. Search strategy to find the best match location of the image matching process to determine the amount of computing of image matching, in the existing image matching method are used to traverse search strategy, it is difficult to reduce the amount of computing. This is a common defect of the existing image matching algorithm. Traditional evolutionary algorithm trapped into the local minimum easily. Therefore, based on a simple evolutionary algorithm and combine the base ideology of orthogonal test then applied it to the population initialization, to prevent local convergence to form a new evolutionary algorithm. Compared the traditional evolutionary algorithm, the new algorithm enlarges the searching space and the complexity is not high. We use this new algorithm in image matching; from the results we reach the conclusion: in the optimization precision and the optimization speed, the new algorithm is efficiency for the image match problem.

In order to solve the problem of slow convergence, an iterative learning control algorithm based on vector plots analysis is proposed. By introducing the geometric principle, that angle and side length of corresponding relationship in the same triangle, and transforming output’s error relationship into the relationship of the angel, the P-type is rewarded or penalized. The new algorithm with angle-correction speeds up significantly the convergence process. For illustration, a numerical simulation is utilized to show its efficiency and superiority. Empirical results show that the new algorithm is faster than P-type to make the system convergence .The method modifying P-type by angle is entirely feasible. The system’s output converges to the desired trajectory. The new algorithm can effectively solve the slow convergence problem. Introduction

Based on rough graph theory, this paper gives a new algorithm in rough network, which generalizes the classical algorithm of exploring maximum flow. This algorithm successfully deals with a kind of complex relationship mining problem between different relationship levels. Simulation shows the effectiveness of this algorithm.

For generating anti-aliasing circle with a high speed and a simple manner, an algorithm which uses only integral operation to convert gray levels of pixels based on modified Bresenham Algorithm is presented. According to the decision parameter established in Bresenham algorithm, the algorithm generated the accurate distances between candidate points and true curve, abandoned the higher-order epsilon with litter effect, and used look-up table or range binary search to calculate the gray levels of the nearest pixels. The new algorithm only modified the local structure of the original algorithm and can generate smooth circle rapidly because of simple structure and only integer operation employed in it. Besides, the new algorithm makes the basic generating algorithm and anti-aliased drawing algorithm can be easily generated under the same framework.

An algorithm of sequence medical images segmen- tation is proposed based on the combination of snakes algorithm and contour interpolation algorithm. Firstly, this algorithm uses snakes algorithm to segment the key layers in which target area change drastically. Then the algorithm calculates the position of reference points in the middle layers with the contour interpolation algorithm. Finally, snakes algorithm is applied again to segment the middle layers. Thus the segmentation of the sequence medical images are accomplished automatically. The experiments showed that the algorithm can obtain the boundary of the desired object from a sequence of medical images quickly and reliably.

According to the time and space distribution of the video sequence image motion vector and the fact that the horizontal component of the non-zero motion vector is often more than the vertical component, this paper presents a cross-diamond search algorithm for motion estimation based on projection from studying the pre-judgment zero motion vector, the starting point of search prediction, the movement type determinant, search strategy formulation and other aspects. The algorithm ensures the image quality basically unchanged. In this case, the search speed of the proposed algorithm speeds up 95-245 times than full search and 4-6 times than the fast motion estimation algorithm. The proposed algorithm has a strong real-time character and an easy hardware implementation.

In this paper, we study the algorithm for solving a kind of new problem called generalized separated continuous conic programming (GSCCP), which has great modeling power. Based on the relationships between GSCCP and its discretization, GSCCP and its dual, we proposed an approximation algorithm which can produce a feasible solution for GSCCP with prescribed precision. We also present a numerical example to illustrate the efficiency of this algorithm. To our knowledge, this is the first algorithm for solving GSCCP so far.

An amended LMS Newton algorithm is proposed which create the nonlinear functional relation of the step factor μ and the error signal e (n). And the next we discuss the factors in the amended algorithm and the performance in the different IP environment. The algorithm is simple and easy to be implemented, Theoretical analysis and computer simulation show that the properties of the algorithm such as convergence speed, and steady state error ate better than those of SVSLMS algorithm.

For the problem of distribution center location about 0-1 mixed integer programming, the necessary and sufficient conditions for the solvable model are given. A heuristic algorithm is designed with the idea of the greedy algorithm by further combined the model characteristics. It is proved that the algorithm could quickly converge to satisfactory solution, and the algorithm is bounded by polynomial time, and the complexity of the algorithm is given.

The traditional power station boiler temperature field reconstruction algorithm is sensitive to the time of flight. In the boiler movement, the temperature field has symmetric distribution feature within the boiler. On the basis of the boiler temperature field reconstruction fundamental by using the acoustic method, the paper presents a new two dimension temperature field reconstruction algorithm, which combines the single path method and genetic algorithm. Firstly, the algorithm makes sure the temperature distribution by using single path function. It uses the points denote the temperatures on each path, and plots the mesh, which can represent the temperature preliminary distribution, by using the Bezier spline principle and linear multistep integration. Finally, the surface mesh is Interpolated and fitted by using genetic algorithm. The experimental result proved that, compared to the least square method, the new reconstruction algorithm has the feature of higher accuracy and higher reconstruction speed.

The EM iterative algorithm is a very general and popular algorithm that commonly used for missing data to find maximum likelihood in recent years, which has the character of stability,flexibility and simlicity.However, the EM algorithm has a great weakness that the convergence speed is slowly; Acceleration of the EM algorithm using the Vector-ε method is proposed in order to solve this problem in this paper.In Multi-user Detection, via this accelerated algorithm, we get a good performance which trends to ML performance and improving the computational efficiency, and compared its speed of convergence with the EM algorithm that Vector-ε acceleration algorithm has faster convergence than the standard EM algorithm.

K-means is a classic algorithm of partition clustering, its speed is very fast, well, the clustering result is very sensitive to the initial cores. As a result, algorithm K-means does not always get the Global Optimization. To solve this problem, this paper proposes a new algorithm partition K-means, which selects the initial cores with a partition method and then cluster the data set with K-means. The experiment shows that the new algorithm gets better effect and better efficiency than the original algorithm K-means.

K-Means algorithm is one of the mostly used foundation algorithm in data mining, it base on a greedy clustering algorithm. This paper will introduce this algorithm and analysis. Then prove the correctness of the algorithm. And then show the productivity of this algorithm. And at last, this paper will show some improvement to K-Means algorithm, including how to choose initial center points, and how to calculate the means. This will improve the algorithm at a certain extent.

An improved genetic algorithm based on hJ1 subdivision is proposed for multimodal optimization problems. With this algorithm, the optimal problems converse to solution of fixed point problems. In this case, whether every individual of the population is a completely labeled simplex can be used as an objective convergence criterion and determined whether the algorithm will be terminated. Finally, a function is used to demonstrate the effectiveness of the algorithm through solving the minimum points distinguished by using the Hessian Matrix.

This paper presents a generating algorithm at pixel level for parametric curve. The parallel particle sub-swarm optimization is used to search the optimal step of curve in forward. Large amount of repeated computing for points are avoided and the result is precise enough. Simulation results show that the parallel method based on particle sub-swarm can be used for searching the optimal step of parametric curve with any degree quickly. At the same time, compared with other methods, this algorithm produces the maximum step efficiently. Since there is no restriction on control point position and curve degree, the algorithm can be extended easily to other parametric curves besides Bézier curves.

This paper presented a high real time target tracking algorithm – mixed difference tracking algorithm MDT. In the proposed algorithm, frame difference and background difference algorithms are combined to get the location of the target. With background difference algorithm the shape of the target can be extracted. Due to the affection of the dynamic background, single background difference algorithm can not get the location of the moving target. To solve this issue the frame difference algorithm is used to estimate the location, and then combine the results of the background difference algorithm and the frame difference algorithm the location and the size of the target can be extracted. And then filtering algorithm is used to remove noise and isolated points. In the experiment it can be seen that the proposed algorithm can tracking object precisely in real time.

The time complexity of B algorithm, one of the intelligent search algorithms, is discussed. By anatomizing some instances, it is pointed out that the cost of calculating the value of heuristic function should be included in the range of time complexity analysis for B algorithm. And then, an algorithm of calculating the value of heuristic function is presented. By analyzing the cost of calculating the value of heuristic function, it is pointed out that the number of recursions in B algorithm is O(n!) in the worst case. Therefore, the time complexity of B algorithm is exponential instead of O(n2).

For fitting of ordered plane data by B-spline curve with the least squares, the genetic algorithm is generally used, accompanying the optimization on both the data parameter values and the knots to result in good robust, but easy to fall into local optimum, and without improved fitting precision by increasing the control points of the curve. So what we have done are: combining the particle swarm optimization algorithm into the B-spline curve fitting, taking full advantage of the distribution characteristic for the data, associating the data parameters with the knots, coding simultaneously the ordered data parameter and the number of the control points of the B-spline curve, proposing a new fitness function, dynamically adjusting the number of the control points for the B-spline curve. Experiments show the proposed particle swarm optimization method is able to adaptively reach the optimum curve much faster with much better accuracy accompanied less control points and less evolution times than the genetic algorithm.