Earticle

In order to detect defects efficiently and improve the quality of products, this paper puts forward the concept about defect classification model and defect association model by a lot of defect data. The technology of similarity is applied to defect classification model, and the idea of Knowledge Discovery in Database is applied to defect association model. Defect classification model can analyze the defect efficiently and provides the basis of solving problems quickly while defect association model can be used to detect early and prevent problem, which can make effective improvements for testing and development. This paper summed up GUI defect model based on a large number of interface defects. The model is useful to improve the accuracy of forecast and be used for test planning and implementation through the practice of several projects.

With the advent of the smart media era, mobile news applications have started to gain attention as a new method of news consumption. Although initial mobile news was nothing but a mobile version of old media outlets, SNS has attracted attention in recent years as a news delivery medium thanks to the massive spread of smart devices and the diffusion of SNS. SNS news service, a horizontal network of journalism driven by prosumers, was able to secure an overwhelming advantage in the speed race. However, this type of news release based on the premise of public consensus may cause errors in judgment in the process of news consumption due to things such as herd mentality, etc.. SNS news services such as ‘Yolk News’ launched recently in an attempt to build a device that can guarantee reliability, and are expected to develop into a form in the future where consumers can make judgments independently by putting various pieces of news together from different angles. Old media will take time to screen the news for their platform, but they are excellent in service quality and objectivity since they are produced by professional groups pursuing journalism. However, they are also subject to the risk of false reporting and loss of confidence triggered by the speed competition and are now losing ground as SNS news services have recently gained confidence beyond expectation. This suggests that the old media should give up the speed competition with SNS news services and need to evolve in a direction to secure objectivity and reliability of the news while maintaining a suitable level of news delivery speed. In recent years, there has appeared a new form of news purely focusing on the value of the news, destroying both the speed and the objectivity of the news. Applications such as New Square are attracting attention as a new form of news-on-demand by providing news in a storytelling manner, allowing readers to be able to identify the sequence and context of events at a glance. As described above, SNS news services have a structural limitation in that the quality and the objectivity cannot be guaranteed because non-professional individuals become the first news providers, while old media have already proven that they cannot keep up with the SNS news services in terms of speed. Therefore, in the smart media era, SNS news services, old media, and news-on-demand will compete and grow by satisfying the speed, objectivity, and consumer value, and the news media ecosystem is likely to change rapidly as these news consuming platforms change.

According to the characters of abrupt change and gradual change, we put forward a self-adaptive method of shot boundary detection, which is based on neighbor frame difference and non-neighbor frame difference. By computing neighbor frame difference, this method can effectively detect abrupt changes. By computing non-neighbor frame difference, gradual changes can be effectively detected. Integrating the two methods, it can almost detect all shot boundaries no matter what the shot change type is. Using this method, we does not need to set global threshold and has high adaptability with low computation cost. Experimental results show it is adaptive and efficient.

As time went by, the convenience derived from information technology and the popularity of shopping online make consumers adapt to online group buying or even addicted in it. However, the explosive growth of group buying market brought problems because of information asymmetry. The group buying websites emerge as bamboo shoots after a spring rain while the numerous group buying websites confused consumers. Therefore, proper evaluation of group buying websites is necessary. The paper proposed AHP-TOPSIS method, a combination of AHP and TOPSIS, to solve this problem under multi-attribute decision-making background. This method takes experts’ opinions into account, both subjective and objective. And the method performs well in solving evaluating problems under multiple attribute decision background. Finally, the experimental results of this method on three group buying websites evaluation proved its good effectiveness and feasibility.

The recovering trust with service failure is very important in business area. Service failure could be causes of incomplete service or employee’s improper response. Although the recovering trust is important in smartphone application, few studies examine how recovering trust influence according to gender. This study investigates the process of recovering trust. The study conducted survey to verify the proposed research based on the theory such as justice, trust, gender etc. The result of the study reveals that the process of recovering trust varies depending on the gender which is one of the key factors that determines users’ behavior.

The reliability of power supply in the distribution network, associated to the normality and effectiveness of electric power supply, plays a significant role in all walks of life, such as urban, rural life and public facilities. Considering about this, it is necessary to evaluate the reliability of electric power supply in the distribution network. A comprehensive evaluation index system for the reliability of distribution network is set up in this paper. Meanwhile, a reliability comprehensive evaluation method is proposed based on matter-element extension model, which can realize the reliability evaluation from the perspective of the system operation. The effectiveness and practicality of the model is verified through the case study of a distribution network.

There are different applications in wireless sensor networks, where sensors with different capacities are deployed. So designing a reliable and scalable routing protocol for a real world wireless heterogeneous sensor network is an interesting task. In this paper, we propose a novel power aware routing for wireless heterogeneous sensor network, which can provide loop less, stateless routing without using any prior neighborhood information. It utilizes both symmetric and asymmetric links to forward a data from source to sink. This protocol explores the neighbor node relations and chooses the best-hop based on node power. It eliminates the low-powered nodes to become hop-nodes in the contention process. It also prolongs the network lifetime. Analysis shows that this scheme significantly outperforms the existing protocol in wireless heterogeneous sensor networks.

A novel field line detection method is proposed based on the directional information of field line. First, the special character of modern soccer playfield is considered, and a novel soccer field model is introduced which includes offside assistant lines. By designing a directional filter, these field lines can be accurately detected. Moreover, the field line detection and tracking for a clip is developed, which can further improve the performance of field line detection. In this paper, we integrate the basic theories of multimedia analysis, pattern recognition and computer vision into content analysis for sports video analysis. Through exploring several components such as object detection, camera calibration and tactic analysis, a systematical system is built up. The proposed algorithms can be not only used for soccer video analysis but also used for other sports video analysis. It provides a promising method for general multimedia analysis.

The paper by analyzing the existence current situation of multiple independent information systems for maritime departments, proposes can use the Service Oriented Architecture (SOA) for the AIS information sharing to do theoretical direction and functional design, in order to solve information island phenomenon of data exchange difficulty between maritime each independent information system. Based on SOA AIS information sharing adopts software engineering method based on service oriented theory, encapsulates AIS data distribution function module as Web service, achieves conveniently called by desktop, mobile phone and so on many kinds of client cross platform, cross platform, cross language, and has the ability to publish the service aggregation integrated with other server GIS service.

This paper presents a new noise model and its application to Bayer color filter array (CFA). We investigate the effect of various noise models. In addition, we study three possible Bayer patterns. For example, current Bayer CFA utilizes RGGB pattern. However, one may consider other color combinations such as RRGB or RGBB patterns. To prove RGGB pattern’s superiority, we investigate each pattern in various noise model and noise level. We also studied the method how to generate noisy image from noisy free image. Traditional noise models do not take into account white balance, gamma and tone correction issues. Simulation results indicate that the new noise model is more reasonable than the conventional models.

In this paper, an interactive color natural image segmentation method is proposed. The method extends the Gaussian Mixture Model (GMM) by taking into account user markers as useful spatial prior. In this way, a distribution combining with color and spatial location is obtained. The distribution is incorporated in a Bayesian MAP approach. The approach is formalized as an iterative energy minimization problem. A direct global minimization technique based on variational method is employed to get binary solution. After each iteration, the largest connected region from foreground is used to update foreground GMMs and achieve more superior performance than original GrabCut method. Extensive experiments are performed on public benchmark datasets such as GrabCut benchmark, Berkeley segmentation database and Graz benchmark. The results show that the proposed method can extract the object region from the complex background accurately.

A person sleeps a third of his or her life, which means that a person sleeps seven or eight hours a day. However, many people suffer from disturbed sleep. In this paper, we try to segment a video into shots using the method of Bhattacharyya Distance. After analyzing these segmented shots, we decide a movement during sleep. We also try to record all the breath from the beginning of the sleep to the ending of the sleep. In the recorded voice, we extract a part of sleep apnea and a part of heavy panting. Sleep apnea is a phenomenon wherein breath stops from 30 seconds to 1 minute and 30 seconds while sleeping. This phenomenon has been known to kill a person from time to time. This sleep apnea causes abrupt heavy panting after no breathing over a period of time. We expect that our research will pave the road to efficiently and effectively judging sleep apnea without specific equipment.

For the shortcoming of smart home monitoring system(SHMS), such as high spend of wiring, difficulty of maintenance, poor penetration using traditional communication etc.., designed a SHMS based on embedded system. By using LPC2000 series processor as a core, selecting power line carrier communication (PLCC) module for interior communication mode, GSM for remote communication mode, and then transplanting RTOS μC/OS-II, the system can monitor environment, security of home. The results indicate that SHMS has the advantages of convenient installation, low cost, easy to upgrade, can realize the local and remote control.

The performance of binarization algorithms is determined by the selection of threshold value for binarization, and most of the previous binarization algorithms analyze the intensity distribution of the original images by using the histogram and determine the threshold value using the mean value of intensity or the intensity value corresponding to the valley of the histogram. The previous algorithms could not get the proper threshold value in the case that doesn't show the bimodal characteristic in the intensity histogram or for the case that tries to separate the feature area from the original image. Therefore, this paper proposed a novel algorithm for image binarization, which, first segments the intensity range of grayscale images to several intervals and calculates a mean value of intensity for each interval, and then repeats the interval integration until getting the final threshold value. The interval integration of two neighborhood intervals calculates the ratio of the distances between mean value and adjacent boundary value of two intervals and determine as the threshold value of the new integrated interval the intensity value that divides the distance between mean values of two intervals according to the ratio. The experiment for performance evaluation showed that the proposed algorithm generates the more effective threshold value than the previous algorithms.

Static hand gesture recognition is very important for human-computer interaction systems, which is widely used in human-computer interaction systems. Using visible RGB images as the input method would be more stable, but the illumination and skin color are big problems; by contrast the depth images is a better choice. A novel algorithm which recognizes static hand gesture based on depth images using 3d shape context feature and improves the performance by arm major axis correction and contour-center sampling is proposed. Arm major axis correction can solve the rotation problem and the sampling, which increases the recognition rate. Besides, using 3d space information makes the algorithm more stable. Experimental results show that the average recognition rate gets to 95.8%, and the performance and speed are superior to existing algorithms. Recognition results can be widely used in the follow-on real scene human-computer interaction (HCI) operations.

Due to the huge information, short length and noise data, the traditional method has poor effect on micro-blog entity relationship modeling. In this paper, a new micro-blog user interests discovering approach based on tag is presented to improve the entity relationship modeling. First the matrix of user tag built by traditional way may generate the problem of sparse matrix in tag recommendation, so we introduce the information of micro-blog and establish the bipartite graph of User-Tag and Tag-Word respectively, then use them to recommend tag to micro-blog users. Meanwhile interactive relationship between users also show their interests, we establish a graph of tag relation by users’ relationship and propose a method called Tag Rank on the basis of this graph to improve the precision of the model. Finally, we combine the two methods to discover user interests. In the experiment, we use several measurement metrics: F-value, precision and the recall rate. It is proven that the new approach in the paper have a perfect performance.

EDZL (Earliest Deadline Zero Laxity) scheduling is known to be at least as EDF (Earliest Deadline First) in task scheduling on symmetric multiprocessor real-time systems; however, there are few works on energy conversation on the EDZL. This paper proposes an on-line Dynamic Voltage Scaling (DVS) algorithm of the global EDZL to reduce energy consumption of real-time tasks. The proposed algorithm dynamically adjusts processor speed at each scheduling point with re-assigning deadlines of active jobs that reduces power consumption of processors while making all real-time tasks schedulable by EDZL. Extensive simulations show that the proposed algorithm reduces power consumption more than the previous algorithm for EDZL.

In common time series analysis methods, the prediction accuracy of the low-order model is poor, and the high-order model is difficult to calculate. Therefore, in this paper, we improve the construction process of the Kalman filtering model, and apply it into time series analysis. The concrete implementation for the improved method is to construct the low-order model with the ARMA method and intercept sufficient delay states, to deduce the state equation and measurement equation of the Kalman filtering model. As the experimental results show that,the improved Kalman filtering model can not only simplify the derivation of the state equation and measurement equation, but also achieve ideal prediction accuracy, the largest prediction error of the experimental data is -0.15%.

The market globalization and the firms’ internationalization hinder the matching of the top managers’ agenda, making it difficult to meet in the same space or time. On the one hand, the appearance of Ubiquitous Group Decision Support Systems (UbiGDSS) ena-bled individuals to gather and make decisions in different spaces at different times, but on the other hand, originated problems related to the lack of human interaction. To under-stand how the arguments used can influence each of the decision-makers, what is their satisfaction regarding the decision made, and other affective issues such as emotions and mood, are some examples of that lack. In order to try to overcome this lack, we propose a theoretical model that is specially designed for agents, helping to understand the interac-tions impact on each agent and their satisfaction with the decision made.

Nowadays, the telecommunication industries are facing substantial competition among the providers in order to capture new customers. Many providers have faced a loss of profitability due to the existing customers migrating to other providers. Customer retention program is one of the main strategies adopted in order to keep customers loyal to their provider. However, it requires a high cost and therefore the best strategy that companies could practice is to focus on identifying the customers that have the potential to churn at an early stage. The limited amount of research on investigating customer churn using machine learning techniques has lead this research to explore the potential of an artificial neural network to improve customer churn prediction. The research proposes Multilayer Perceptron (MLP) neural network approach to predict customer churn in one of the leading Malaysian’s telecommunication companies. The results are compared against the most popular churn prediction techniques such as Multiple Regression Analysis and Logistic Regression Analysis. The result has proven the supremacy of neural network (91.28% of prediction accuracy) over the statistical models in prediction tasks. Overall, the findings suggest that a neural network learning algorithm could offer a viable alternative to statistical predictive approaches in customer churn prediction.

The noise in the acoustic emission (AE) signal must be removed to identify the mode of AE signal accurately. The Wavelet threshold de-noising method shows some unique advantages. Based on the threshold selection risky problem, K-means clustering method was used to classify the high-frequency coefficients by the wavelet decomposition to determine the removal threshold for the wavelet coefficients corresponding to the noise, and achieve the de-noising purpose. Hard-threshold method and soft-threshold method were applied to AE signal through the wavelet threshold de-noising. The thresholds generated by K-means clustering approach and the Donoho method improved were respectively used as the threshold for the de-noising of the wavelet coefficients. The experimental results show that the method proposed is superior to the Donoho method improved in the three indicators of signal to noise ratio, root mean square error.

This paper proposes a novel 3D wrist gesture recognition method as the human-machine interface for mobile robot control. A sequence of depth maps was extracted from Kinect sensor on wrist gestures of an operator and a wrist gesture was specified by a sequence of 24-directional codes, which represents 3D motions with feature values of a single type, including directional and depth information. HMM algorithm was used to recognize wrist gestures, in which a fully connected hidden Markov model was designed and trained by Viterbi method and Baum-Welch method and the forward-backward procedure was used for the likelihood evaluation. The performance evaluation considering various factors showed above 96% of recognition rate and higher recognition rate on wrist gestures with complex motions.

In order to distinguish the sharpening operation in a manipulated image, a sharpen forgeries detection algorithm is proposed. Combined with efficient image signal processing method, a new non-subsampled dyadic contourlet transform based on contrast à trous wavelet is designed. In contourlet domain, sharpening characteristics are analyzed by multi-scale and multi-directional methods. A classifier is trained by artificial intelligence methods to complete sharpening tampered image detection. Experimental results show that the algorithm in this paper can detect possible sharpening and locate the tampering boundary accurately.

In order to demonstrate the effectiveness of sparse representation techniques for speaker recognition, a dictionary of feature vectors belonging to all speakers is constructed by total variability i-vectors. Each feature vector from unknown utterance is expressed as linear weighted sum of a dictionary. The weights are calculated using Block Sparse Bayesian Learning (BSBL) where the sparsest solution can be obtained. By exploiting the speech signal’s block structure and intra-block correlation, the system performance is improved. The experimental results validate that our method outperforms the baseline systems and the system using Orthogonal Matching Pursuit (OMP) algorithm on the typical corpus and realizes the identity validation function.

GPS devices typically make use of images that are excessively large to be stored as a Bit Map, so the images are always compressed. The widespread compression technique is JPEG; however JPEG has a disadvantage – JPEG assumes that the average color of beginning of each blocks line is commonly similar to the average color of end of its preceding blocks line. Almost always this assumption is wrong for GPS images. This paper proposes a method to adjust JPEG order of compression to an improved order that is more suitable for GPS images.

In order to explore how to design an information system with ideal network property under the background of information era, contrastive analysis is conducted in this paper on the performances of “tree” network and fully connected network. The mathematical models of “tree” information system and fully connected network are established, of which the network performance indicators are calculated by complex network theory and analyzed by comparison. As only partial indicators of both networks are favorable, the plan is proposed that information system is established through optimization based on both of the two network structures. The new algorithm to build information system network model proposed in this paper and the plan that information system is established are of significant reference value for relevant fields.

The open structure of MANET, dynamic topology, and infrastructure mobile nodes constituting are threated from many types of attacks. Therefore, what increases the reliability between the mobile nodes and provides efficient routing is an important part to influence network performance. Attack by malicious nodes presenting within the network destructs communication path or discards the packet and the damage is very large. However, it is not easy to detect attacks using these properties. In this paper, we propose an authentication technique to provide secure communication through secure routing scheme for safe data transmission between the nodes and exhaustive reliability testing. Cluster structure is used for authentication technique of the proposed technique and cluster head acts as a certificate authority and is managed authentication information of member nodes. The performance of the proposed technique was confirmed by experiments.

For path planning of mobile robot, the traditional Q learning algorithm easy to fall into local optimum, slow convergence etc. issues, this paper proposes a new greedy strategy, multi-target searching of Q learning algorithm. Don't need to create the environment model, the mobile robot from a single-target searching transform into multi-target searching an unknown environment, firstly, by the dynamic greedy strategy exploring interim to use unknown environment, improve learning ability that mobile robot learn the environment, improve the convergence of the mobile robot speed. And a large number of improved Q-learning algorithms are applied to mobile robot optimization simulation in unknown environment, by comparing with traditional Q algorithm, theory and experiment proved that improved Q-learning algorithm speed up the convergence rate of the robot, improve collision avoidance capability and learning efficiency.