Applied Mechanics and Materials Vols. 263-266

Abstract: The dynamical evolutionary algorithm (DEA) is a new evolutionary algorithm based on the theory of statistical mechanics, however, DEA converges slowly and often converge at local optima for some function optimization problems. In this paper, a hybrid dynamical evolutionary algorithm (HDEA) with multi-parent crossover and differential evolution mutation is proposed for accelerating convergence velocity and easily escaping suboptimal solutions. Moreover, the population of HDEA is initialized by chaos. In order to confirm the effectiveness of our algorithm, HDEA is applied to solve the typical numerical function minimization problems. The computational complexity of HDEA is analyzed, and the experimental results show that HDEA outperforms the DEA in the aspect of convergence velocity and precision, even the two algorithms have the similar time complexity.

Abstract: There exist principles of Relevancy, Orthogonality and Domain Adaptation for auxiliary problems (APs) selection in Alternating Structure Optimization (ASO) algorithm. Convex Alternating Structure Optimization (cASO) algorithm is an improved one based on ASO algorithm, whose kernel still lies in creating excellent APs. In order to validate the effectiveness of the preceding principles in cASO algorithm, many types of APs were created by taking example of Chinese syntactic chunking. Experimental results and analyses both demonstrate that those principles still hold in cASO algorithm.

Abstract: In order to detect the boundary points of clustering efficiently, we proposed a novel algorithm which combined grid technology and gradient operator. In this algorithm the grid technology is used to enhance the speed of the data processing, and Prewitt gradient operator is applied to calculate gradient in 3×3 grid region from eight directions, the maximum being the central grid gradient. The gradient is used to judge whether the grid is the boundary grid or not, and a point in the boundary grid is a boundary point. Putting the method of image boundary processing into the practice of processing cluster boundary is a fresh idea for the research on cluster boundary. The experimental results indicate that this algorithm can effectively detect boundary of clusters in datasets with noises/outliers and have high running efficiency.

Abstract: In this paper, a video summarization representation algorithm was proposed in compressed domain. In particular, Rough sets(RS) theory is introduced for video analysis to increase. Firstly, DCT coefficients and DC coefficients are extracted from video image sequences, so an Information System can construct with DC coefficients. Then Information System is reduced by ruduction theory of RS, the representation of the video frame is obtained by reduced DC coefficients. Finally, we can obtain the reduced Information System, i.e. the Core of Information System. Since the Core contained all the information in video sequences, and at the same time it banished redundant video frame, so it can be viewed as the effective summarization representation. Experimental results indicate that the algorithm can efficiently generate a set of summarization representative of videos sequences and enjoys following advantages. Only a subset of video frames considered during video analysis, so it can avoid the computational complexity, the video summarization representation becomes more scientific than previous methods.

Abstract: A novel textual integrity detection algorithm based on matrix coding is presented in this paper, which can improve the embedding efficiency and security of digital watermark in the WORD document. The experimental results show that better invisibility and security of watermark can be obtained by using matrix coding.

Abstract: Nature image has complex structure and texture. In past years, people copy the pixels, which come from the surrounding neighborhood or selected example, into damaged region along isophate direction. This might cause such result as blur edges, which looks “unreasonable” to human eyes. This paper proposes a novel algorithm which simultaneously inpaints structures and textures of damaged images. This algorithm use TV to decompose the image into two parts. We inpaint the cartoon image part firstly by boundary reconstruction. Then we do texture synthesis to texture image part guided by boundary reconstruction. The method aims at inpainting structure and texture simultaneously and produces good results for texture with complex structure.

Abstract: With the development of visual measurement system , the visual measurement system of single camera which applied in the imaging theory of optical feature points , it has been widespread used in the modern production . Due to the limit of the environment in scene , the visual measurement system of single camera could not measure the shield between the measured objects each other . Focus on this problem , present a kind of the the knowledge of measurement network based on the visual measurement of single camera , set up the measurement network system via the multi-control points . Measure the optical feature points in every network control point via the visual measurement system of single camera; Normalize the measured coordinates into the same world coordinate system , then get the global data , and achieve the overall measurement ; Apply in the three coordinate measuring machine(CMM) to have the simulation experiment , the result indicates that the maximum absolute tolerance between the normalization coordinates via the center of point set and the ones measured by CMM directly is 0.058mm ; But the one between the coordinate repeatedly measured at the position of one network control point is 0.066mm ; So the data normalization has the advantage of more high precision .

Abstract: Video object tracking is essential algorithm for computer vision applications. An object tracking algorithm using combining motion constraints model and online multiple instance boost random ferns is proposed, which use IIR filter to obtain online learning for random ferns, and the random ferns are selected by online multiple instance boosting to construct classifier of online multiple instance boost random ferns. To reduce effects of tracking error accumulation, object motion constraint model is constructed to constrain the results classified by online multiple instance boost random ferns to locate object correctly, and construct positive and negative set to online update the classifier. The experiment shows that the proposed method achieves competitive detection results, which are comparable with state-of-the-art methods.

Abstract: Close-range photogrammetry is a technique of calculating the location, size and shape of measured object by photography whose object distance is generally not greater than 300 meters. Three-dimensional (3D) model reconstruction based on close-range photogrammetry has higher efficiency than that based on Light Detection And Ranging (LiDAR) technique since acquiring texture data simultaneously. This technology reduces the consuming time of 3D model reconstruction, while ensuring high precision. In this paper, processes and key technologies of 3D model reconstruction based on portable close-range photogrammetry are provided, and it feasibility of the technology is verified via taking Taizhou TV Tower as an example.

Abstract: This paper presents the developing of a low-cost eye-tracking system by modifying the commercial-over-the-shelf camera to integrate with the proper-tuned open source drivers and the user-defined application programs. The system configuration is proposed and the gaze-tracking approximated by the least square polynomial mapping is described. Comparisons between other low-cost systems as well as commercial system are provided. Our system obtained the highest image capturing rate of 180 frames per second, and the ISO 9241-Part 9 test performance favored our system, in terms of Response time and Correct response rate. Currently, we are developing gaze-tracking accuracy application. The real time gaze-tracking and the Head Movement Estimation are the issues in future work.