Applied Mechanics and Materials Vols. 599-601

Abstract: A new neighborhood variable precision rough set modal is presented in this paper. The modal possesses the characteristics of neighborhood rough set and variable precision rough set, so it can overcome shortcomings of classic rough set which only be fit for discrete variables and sensitive to noise. Based on giving the definitions of approximate reduction, lower and upper approximate reduction, lower and upper distribution reduction, two kinds of algorithms to confirm lower and upper distribution reduction were advanced. The modal was applied to diagnose one frequency modulated water pump vibration faults. The result shows the modal is more suitable to engineering problems, because it can not only deal with continues variables but also be robust to noise.

Abstract: A novel blind source separation (BSS) algorithm for linear mixture signals is proposed. It is shown that the property can be used to separate source signals by finding an un-mixing matrix that maximizes the cost function value of separated signals. Simulation results illustrate the efficiency and the good performance of the algorithm.

Abstract: In order to improve the accuracy of correlated noise (CN) model for distributed video coding (DVC), this paper proposes a novel distribution parameter fitting algorithm based on the minimum Euclidean distance. The presented method can obtain the final fitted distribution parameter by using the minimum Euclidean distance to compare the Laplace probability density function (PDF) with the PDF computed utilizing the actual residual frame data. Experiment results show that the proposed distribution parameter fitting algorithm can improve the rate-distortion (R-D) performance of DVC significantly.

Abstract: According to the characteristics of interferogram, we design a new spike detection method, which firstly filters an interferogram with two different ways and then detects spikes based on both results. Theoretical analysis and computer simulation shows that this algorithm performs well in detecting spikes in any position of an interferogram with high accuracy, and can be easily implemented in hardware.

Abstract: The community structure is a vital property of complex networks. As special networks the weighted networks also have community structure. Nowadays the studies of overlapping community draw attentions of researchers. However, the scale of networks become huge, so it requires the algorithm has lower time complexity and higher classification accuracy. Many existing algorithms cannot meet these two requirements at the same time. So we propose a novel overlapping community detection algorithm. Firstly we apply maximum degree node and its some special adjacent nodes as the initial community, and then expand the initial community by adding eligible nodes to it, finally other communities can be found by repeating these two steps. Experiments results show that our algorithm can detect overlapping community structure from weighted networks successfully, and also reveal that our method has higher division accuracy and lower time complexity than many previously proposed methods.

Abstract: The measure of similarity is an essential basic work in the field of pattern recognition, machine learning and data mining, etc. The shape similarity distance (SSD) which considering vector difference factors in the calculation of similarity, is an amendment of the classical distance. Based on the classical Euclidean distance, Manhattan distance and the SSD, this paper completed the clustering on the multiple datasets that contained shape features data of objects. The experimental results indicated that compared with the classical distances, the SSD can consider the characteristics of shape of theses objects to estimate their similarity.

Abstract: -Aiming at getting a high efficient network routing decision strategy, to settle problems of slow convergence speed and easily to fall into local optimal, this paper proposes a new decision strategy based on ant colony algorithm. Memory device to record each time pheromone value and pheromone differences value of the adjacent times to decide follow the former route or find a new one are the focus of this paper. The new decision technology accelerates the convergence rate, improves network utilization rate and accuracy of network routing.

Abstract: In this paper, we propose a improved error concealment technique based on multi-view video coding to recover damaged video images. At first,It uses BMA(Boundary Matching Algorithm) method to recover the lost or erroneously received motion vector or disparity vector,then combining inter-view correlation, temporal correlation and spatial correlation to recover the lost blocks. The JM12.0 model of H.264 standard is used to evaluate the algorithm. And the experimental results show that our algorithm achieved a better image reconstruction.

Abstract: An improved frame loss concealment algorithm based on the Mixed Excitation Linear Prediction (MELP) is presented in this paper. It introduces the Future frame to recover the lost frame and the intermediate frame (U/V frame and V/U frame) to determine the type of lost frame more accurately. Meanwhile, it proposes the Lagrange polynomial approach algorithm and the linear prediction algorithm to recover the parameters of lost frames in different frame types. The PESQ-MOS test shows that the synthetic speech quality has been improved with the algorithm which this paper has proposed.

Abstract: Traditional imaging method use high precision mirrors to focus the lightfield toobtain object’s imaging, this method is affected by aberrations and atmosphere turbulence. However lensless imaging method can get high resolution without using lens and mirrors ,it will make the system compact light weight. In this paper,we analysis and simulated a new lensless imaging method with intensity correlation, by this way,we can get high resolution imaging by bare Charge Coupled Device (CCD) only,and the simulations show that this method can obtain good image to low angular diameter object.