Search

We propose a novel generic inverted index framework on the GPU (called GENIE), aiming to reduce the programming complexity of the GPU for parallel similarity search of different data types. Not every data type and similarity measure are supported by GENIE, but many popular ones are. We present the system design of GENIE, and demonstrate similarity search with GENIE on several data types along with a theoretical analysis of search results. A new concept of locality sensitive hashing (LSH) named tau-ANN search, and a novel data structure c-PQ on the GPU are also proposed for achieving this purpose. Extensive experiments on different real-life datasets demonstrate the efficiency and effectiveness of our framework. The implemented system has been released as open source: https://github.com/SeSaMe-NUS/genie.

Saliency prediction typically relies on multiple features that are combined in the ways of weighted summation or multiplication to form a saliency map, which is heuristic and hard for generalization. In this paper, a novel multiple graph label propagation integration framework for saliency object detection algorithm is proposed. The proposed algorithm is divided into four steps. First, an input image is segmented into superpixels which are represented as nodes in a graph and transformed from RGB color space into CIE L*a*b* color space. Second, combined by texture features, we measure the similarity of two adjacent superpixels for each feature, which is represented as an affinity matrix. Then, to generate the salient seeds, we adopt the color boosting Harris points as salient points to catch the corners or marginal points of visual salient region in color image. The saliency points provide us a coarse location of the salient areas. In the last step, the graphs are combined into label propagation framework to obtain the saliency maps. We propose efficient optimization algorithms for the proposed approach, which generate sparse weighted coefficients that allow identifying the graphs which are important or not for salient object detection easily. Experiments on four benchmark databases demonstrate the proposed method performs well when it violates the state-of-the-art methods in terms of accuracy and robustness.

Natural image segmentation is often a crucial first step for high-level image understanding, significantly reducing the complexity of content analysis of images. LRAC may have some disadvantages. (1) Segmentation results heavily depend on the initial contour selection which is a very skillful task. (2) In some situations, manual interactions are infeasible. To overcome these shortcomings, we propose a novel model for unsupervised segmentation of viewer’s attention object from natural images based on localizing region-based active model (LRAC). With aid of the color boosting Harris detector and the core saliency map, we get the salient object edge points. Then, these points are employed as the seeds of initial convex hull. Finally, this convex hull is improved by the edge-preserving filter to generate the initial contour for our automatic object segmentation system. In contrast with localizing region-based active contours that require considerable user interaction, the proposed method does not require it; that is, the segmentation task is fulfilled in a fully automatic manner. Extensive experiments results on a large variety of natural images demonstrate that our algorithm consistently outperforms the popular existing salient object segmentation methods, yielding higher precision and better recall rates. Our framework can reliably and automatically extract the object contour from the complex background.

To evaluate the environmental effectiveness of the control measures for atmospheric pollution in Shijiazhuang of China, a large-scale controlling experiment for emission sources of atmospheric pollutants (i.e., a temporary emission control action, TECA) was designed and implemented during November 1, 2016 to January 9, 2017. Under the unfavorably meteorological conditions, the mean concentrations of PM2.5, PM10, SO2, NO2, and chemical species (Si, Al, Ca2+, Mg2+) in PM2.5 during the control action and heating period (CAHP) still decreased by 15 %, 26 %, 5 %, 19 %, 30.3 %, 4.5 %, 47.0 % and 45.2 %, respectively, compared to the no control action and heating period (NCAHP); indicating that the control measures of atmospheric pollution in Shijiazhuang were effective and was in a right direction. Overall, the effects of control measures in suburbs were better than those in urban area, especially for the control effects of particulate matters sources. The control effects for emission sources of carbon monoxide (CO) were not apparent during the TECA period, especially in suburbs, which is likely due to the increasing usage of domestic coal in suburbs along with the temperature decreasing. The results of PMF analysis showed that crustal dust, secondary sources, vehicle emissions, coal combustion and industrial emissions were major PM2.5 sources. Compared to the whole year, the contributions of coal combustion to PM2.5 increased significantly during the CAHP and after control action (ACA); while the contribution proportions of crustal dust and vehicle emissions to PM2.5 decreased apparently during the CAHP. The contribution concentrations and proportions of crustal dust and vehicle emissions to PM2.5 during the CAHP also decreased significantly compared to ACA. The pollutant's emission sources during the CAHP were in effective control, especially for crustal dust and vehicles. While the necessary coal heating for cold winter and the unfavorably meteorological conditions had an offset effect on the control measures for emission sources to some extent. Meanwhile, our results also illustrated that the discharge of pollutants was still enormous even under such strict control measures. The backward trajectory and potential source contribution function (PSCF) analysis in the light of atmospheric pollutants suggested that the potential sources-areas mainly concentrated in surrounding regions of Shijiazhuang, i.e., south of Hebei, north of Henan and Shanxi. The regional nature of the atmospheric pollution in Northern China Plain revealed that there is an urgent need for making cross-boundary control policy except for local control-measures given the high background level of pollutants. The TECA is an important practical exercise but it can't be advocated as the normalized control measures for atmospheric pollution in China. The direct cause of atmospheric pollution in China is the emission of pollutants exceeds the air environment's self-purification capacity, and the essential reason is unreasonable and unhealthy pattern for economic development of China.

Without sufficient preparation and on-site management, the mass-scale unexpected huge human crowd is a serious threat to public safety. A recent impressive tragedy is the 2014 Shanghai Stampede, where 36 people died and 49 were injured in celebration of the New Year’s Eve on December 31st 2014 in the Shanghai Bund. Due to the innately stochastic and complicated individual movement, it is not easy to predict collective gatherings, which can potentially leads to crowd events. In this chapter, with leveraging the big data generated on Baidu Maps, we propose a novel approach to early warning such potential crowd disasters, which has profound public benefits. An insightful observation is that, with the prevalence and convenience of web map service, users usually search on the Baidu Maps to plan a routine, which reveals the users’ future destinations. Therefore, aggregating users’ query data on Baidu Maps can obtain priori and indication information for estimating future human population in a specific area ahead of time. Our careful analysis and deep investigation on the Baidu Maps’ data on various events also demonstrate a strong correlation pattern between the number of map query and the number of positioning in an area. Based on such observation, we propose a decision method utilizing query data on Baidu Maps to invoke warnings for potential crowd events about 1~3 h in advance. Then we also construct a machine learning model with heterogeneous data (such as query data and positioning data) to quantitatively measure the risk of the potential crowd disasters. We evaluate the effectiveness of our methods on the data of Baidu Maps.

Laser stripe center extraction is a key step for the profile measurement of line structured light sensors (LSLS). To accurately obtain the center coordinates at sub-pixel level, an improved gray-gravity method (IGGM) was proposed. Firstly, the center points of the stripe were computed using the gray-gravity method (GGM) for all columns of the image. By fitting these points using the moving least squares algorithm, the tangential vector, the normal vector and the radius of curvature can be robustly obtained. One rectangular region could be defined around each of the center points. Its two sides that are parallel to the tangential vector could alter their lengths according to the radius of the curvature. After that, the coordinate for each center point was recalculated within the rectangular region and in the direction of the normal vector. The center uncertainty was also analyzed based on the Monte Carlo method. The obtained experimental results indicate that the IGGM is suitable for both the smooth stripes and the ones with sharp corners. The high accuracy center points can be obtained at a relatively low computation cost. The measured results of the stairs and the screw surface further demonstrate the effectiveness of the method.

In this study, a novel saliency detection algorithm based on linear neighbourhood propagation is proposed. The proposed algorithm is divided into three steps. First, the authors segment an input image into superpixels which are represented as the nodes in a graph. The weight matrix of the graph, which indicates the similarities between the nodes, is calculated by linear neighbourhood reconstruction. Second, the nodes, which are located at top, bottom, left and right of image boundary, are labelled as boundary priors. Then, based on weight matrix, label propagation is used to propagate the labels to unlabelled nodes. They rank the nodes according to the label information and select the nodes with minor information as saliency priors. Last, based on saliency priors, saliency detection is carried out by label propagation again. The nodes with more information are considered as saliency regions. Experimental results on three benchmark databases demonstrate the proposed method performs well when it is against the state-of-the-art methods in terms of accuracy and robustness.

Creating an efficient and economic trip plan is the most annoying job for a backpack traveler. Although travel agency can provide some predefined itineraries, they are not tailored for each specific customer. Previous efforts address the problem by providing an automatic itinerary planning service, which organizes the points-of-interests (POIs) into a customized itinerary. Because the search space of all possible itineraries is too costly to fully explore, to simplify the complexity, most work assume that user's trip is limited to some important POIs and will complete within one day. To address the above limitation, in this paper, we design a more general itinerary planning service, which generates multiday itineraries for the users. In our service, all POIs are considered and ranked based on the users' preference. The problem of searching the optimal itinerary is a team orienteering problem (TOP), a well-known NP-complete problem. To reduce the processing cost, a two-stage planning scheme is proposed. In its preprocessing stage, single-day itineraries are precomputed via the MapReduce jobs. In its online stage, an approximate search algorithm is used to combine the single day itineraries. In this way, we transfer the TOP problem with no polynomial approximation into another NP-complete problem (set-packing problem) with good approximate algorithms. Experiments on real data sets show that our approach can generate high-quality itineraries efficiently.

The feasibility of applying rapid and non-destructive analysis based on Visible/ Near-Infrared Spectroscopy (VIS/NIR spectrometer) in detecting Malus Asiatica Nakai’s firmness during different shelf lives was discussed in this manuscript. Based on the best preprocessing technique found for each method, calibration models established by Partial Least-Square (PLS), Principal Component Regression (PCR), and Least Squares-Support Vector Machine (LS-SVM) were compared. The optimal models for firmness of Malus Asiatica Nakai were obtained by PLS with SNV as the best pretreatment. On the 1st day of shelf life, the correlation coefficients of prediction (Rp) and the Root-Mean-Square Errors of Prediction (RMSEP) were 0.7999 and 0.7994, respectively. On the 4th day of shelf life, the indexes were 0.7494 and 0.5856. But with the extension of shelf life, weightlessness rate of Malus Asiatica Nakai was reduced, the mealiness phenomenon became serious, quality had dropped sharply, hence the research of Malus Asiatica Nakai became insignificant. Results showed that, the potential of non-destructive measure for predicting accepted fruit parameters enabled the determination of optimum harvest date and fruit quality in different shelf lives.

The relationship between ecological and socioeconomic systems in ecological function conservation areas (EFCAs) in China is analyzed from a governance perspective. Lashihai watershed in China’s southwestern Yunnan Province was chosen as a case study area, where leaders of 81 villager groups were interviewed through questionnaire surveys and anecdotal evidence was collected from focus group discussions. Our study found that the rehabilitated ecosystems in Lashihai, arising from conservation actions, provided the local communities with increasing natural capital to pursue horseback tourism as an important means of livelihood. Also, bonding social capital, together with unique cultural and ethnic assets, may have been instrumental in the formation of horseback tourism teams in some villager groups. However, the lack of higher-level government involvement and coordination among horseback tourism teams appeared to have led to a situation, where the rapidly developing tourism teams have started to impose threats on the wetland ecosystem. While highlighting some limitations of self-governance in adapting to complex and fast-changing socioeconomic conditions, the study called for the potential importance of fostering adaptive co-management to help modify the emerging undesired interconnectedness in social-ecological systems in Lashihai. For the future’s successful governance of integrated ecosystem management in EFCAs in China, the study also made brief discussion on some key elements of the adaptive co-management.

Line structured light sensors (LSLSs) have gained more and more applications in industry. An interested profile can be easily obtained through the analysis of laser-object intersection stripe. But one sensor is inadequate to get a closed crosssection profile due to the obstacle of the laser light. Thus, multiple LSLSs were integrated as a whole for profile inspection and a numerical calibration method was also proposed. Firstly, the laser planes from all laser projectors were adjusted to coincide with the target plane by adjusting the fixtures of the laser projector. For each sensor, origin of the world coordinate system (WCS) was fixed at the center of a corner calibration dot with its X and Y axis coincide with the row and column direction of target dots. Each sensor camera captured one image of the same target. The relationship between the pixel coordinate system (PCS) and the WCS was established using an interpolation method via the world coordinates of target dot centers and their corresponding pixel coordinates. Then the measurement points from all the sensors were transformed into the global WCS, and a closed cross-section profile can be achieved. This proposed method neither need to establish the intrinsic, the extrinsic and the distortion models of the camera, nor need to solve the complex optimization equations to determine the model coefficients. Finally, a workpeice with stairs and a rectangular block were inspected. The comparison with the measuring results from the coordinate measuring machine further validates the high accuracy of the proposed method.

Choosing a good location when opening a new store is crucial for the future success of a business. Traditional methods include offline manual survey, analytic models based on census data, which are either unable to adapt to the dynamic market or very time consuming. The rapid increase of the availability of big data from various types of mobile devices, such as online query data and offline positioning data, provides us with the possibility to develop automatic and accurate data- driven prediction models for business store site selection. In this paper, we propose a Demand Driven Store Site Selection (DD3S) framework for business store site selection by mining search query data from Baidu Maps. DD3S first detects the spatial-temporal distributions of customer demands on different business services via query data from Baidu Maps, the largest online map search engine in China, and detects the gaps between demand and supply. Then we determine candidate locations via clustering such gaps. In the final stage, we solve the location optimization problem by predicting and ranking the number of customers. We not only deploy supervised regression models to predict the number of customers, but also use learning-to-rank model to directly rank the locations. We evaluate our framework on various types of businesses in real-world cases, and the experiment results demonstrate the effectiveness of our methods. DD3S as the core function for store site selection has already been implemented as a core component of our business analytics platform and could be potentially used by chain store merchants on Baidu Nuomi.

User identification across domains draws lots of research effort in recent years. Although most of existing works focus on user identification in a single space, in this paper, we first try to identify users by fusing their activities in cyber space and physical space, which helps us obtain a comprehensive understanding about users' online behaviours as well as offline visitation. Out profound insight to tackle this problem is that we can build a connection between the cyber space and the physical space with the stable location distribution of IP addresses. Thus, we propose a novel framework for user identification in cyber-physical space, which consists of three key steps: 1) modeling the location distribution of each IP address; 2) computing the co-occurrence with an inverted index to reduce the space and time cost; and 3) a learning-to-rank tactic to fuse user's features shared in both spaces to improve the accuracy. We conduct experiments to identify individual users from mobile query logs (generated in cyber space) and trajectory data (generated in physical space) to demonstrate the efficiency and effectiveness of our framework.

As a non-contact measuring apparatus, line structured light sensor (LSLS) can only get one profile of an object without the combination with other motional axes. To achieve the complete 3D measurement, a rotation-translation platform was integrated with the LSLS, and a cylinder based calibration method was also brought out. Firstly, the calibration model was proposed to determine the transformation matrix between the measuring coordinate frame (MCF) and the sensor coordinate frame (SCF). This model relies on the fact that the projection of an arbitrary intersection profile between the laser plane and the cylinder in its axis direction lies on a circle with a radius equal to the cylinder. Then, for a specified rotated angle and translated position of the object, the measured data from the SCF could all be transformed into the MCF, and the complete surface data could be obtained. Finally, a cylinder and a rectangular block were inspected by the proposed method. The surface data was successfully obtained and their intersection profiles indicate a high measuring accuracy of the proposed method. The method was further verified by the measured results of a screw surface.

With the advantages of easy data processing and fast measuring speed, line structured light sensors (LSLSs) have gained more and more applications. CCD camera is a core component of the sensor. The distortion of its lens will severely degrades the measuring accuracy. To enhance the measuring quality, a numerical calibration method is brought out that is based on linear transformation over triangular domain. Based on the pinhole imaging principle, a linear transformation model was established which is easy to compute the profile’s world coordinates according to its pixel coordinates over each triangular domain. The triangle domains are achieved using Delaunary triangulation via the centers of target dots. The triangle number that each center point of the laser stripe locates is determined by T-search method. A linear approximation error model to the lens distortion is also established and the approximation errors are getting larger when the interval spacing of the calibration dots increases. Measuring results show that the relative error of this proposed method in horizontal and vertical direction can reach 0.0630% and 0.0802%, respectively. The calibration error grows with the increasing of the target’s dot interval that corresponds with the trends of the linear approximation error. This further validates the proposed calibration method.

This paper presents a new design concept for the single treatment of X/O type lower limb rehabilitation instrument, the new design concept, modeling and simulation using SOLIDWORKS, the force analysis using ADAMS, control the use of single-chip microcomputer, the real realization of a set of rehabilitation, massage and promoting blood circulation in one of the functions of the machine, and has a very broad market application prospect.

The stoichiometric composition of plant reproductive organs varies with environmental conditions at a large geographical scale, which has potentially important consequences both for plant reproductive success and for the nutrition of consumers. In the study reported here, we investigated variations in the concentrations of macronutrients in acorns of the oriental oak (Quercus variabilis) in relation to environmental variables across a geographic gradient spanning temperate and subtropical latitudes. Across this gradient, nitrogen (N) and calcium (Ca) concentrations of the acorns had higher coefficients of variance (CV 18–23 %) than the other elements (CV 8–14 %). With increasing latitude (LAT), acorn N, potassium (K), and magnesium (Mg) concentrations increased, whereas those of sulfur (S) and Ca and the carbon:nitrogen (C:N) and S:N ratios decreased. The K and Mg concentrations of the acorns were positively correlated with the concentrations of these elements in the soil, whereas N and Ca concentrations were inversely correlated with soil concentrations. The C:N:P (phosphorus) molar ratio of the acorns was 1,008:17:1. A low temperature and short growing season stimulated seeds to accumulate more macronutrient elements in the north to ensure reproductive success, while high soil Ca in the north led plants to allocate less Ca to their acorns. The pattern of variation in acorn stoichiometric traits along LAT may be the result of interacting effects of plant reproduction strategy and environmental factors. Our results can help improve our understanding of the effects of global changes on the chemical compositions of plant reproductive organs and advance our knowledge of global patterns of ecological stoichiometry in reproductive organs.

Local discriminant embedding (LDE) only utilizes the local information and ignores the nonlocal information. Although linear discriminant analysis (LDA) utilizes the local information and the nonlocal information simultaneously, it treats these two kinds of information equally. As we know, the local information and the nonlocal information are both effective for feature extraction, but they have different roles in feature extraction. To utilize the local information and the nonlocal information simultaneously and utilize them distinctively, a new feature extraction approach called weighted linear embedding (WLE) is proposed by using the Gaussian weighting. Further, a method to set the optimal parameter of the Gaussian weighting is put forward. WLE is evaluated on YALE, FERET face databases, the PolyU palmprint database, and the PolyU finger-knuckle-print database. The experimental results demonstrate the effectiveness of WLE.

The dimensionality of sample is often larger than the number of training samples for highdimensional pattern recognition such as face recognition. Here linear discriminant analysis (LDA) cannot be performed directly because of the singularity of the within-class scatter matrix. This is socalled “small sample size” (SSS) problem. PCA plus LDA (FDA) and Direct LDA (DLDA) are two popular methods to solve the SSS problem of LDA. In this paper, we point out the relationship of these two methods and discuss the deficiency of DLDA. Then a normalized direct linear discriminant analysis (NDLDA) method which overcomes DLDA’s deficiency is proposed. Experiments on ORL, YALE and AR face databases show NDLDA’s superiority over DLDA and FDA.

In order to improve the utilization and competitiveness of shared vehicles, the emerging car sharing system tends to provide one-way mode without reservation and allow remote borrowing. Unbooked one-way vehicle sharing system is characterized by the opening of vehicle mobility, allowing vehicles to return at other stations. But it leads to the imbalance of demand distribution in a certain period of time. When the demand is satisfied and the trip is completed, the vehicle will deviate from the original layout. The subsequent demand for areas with large demand can not be met, and vehicles with low demand are idle. This paper considers the sustainable development of shared car rental companies. In order to optimize the profit of shared car rental enterprises and enhance their competitiveness, intelligent algorithm is used to optimize the scheduling of vehicles with different outlets. So as to maximize service quality and company profits. First, a mathematical model for the scheduling of shared car is established. Secondly, different scheduling strategies are designed for different network scheduling. At last, an artificial fish swarm algorithm is used to analyze the case in MATLAB. There are two car outlets in the car rental company, with a maximum of 20 cars available for lease at each location, and the most profitable scheduling method when the most of the 5 cars are scheduled to be transferred every day.

With the global water crisis is getting more and more serious, industrial and sewage treatment are becoming the urgent problem need to be solved.In this paper, sewage treatment project of a factory is research object, the design using the upper computer configuration software from YAKONG of China which named KINGVIEW6.53 and using the PLC from Siemens Ltd. The automatic control of PLC has been successfully completed in sewage treatment ,which has improved the operation efficiency and management level of sewage treatment equipment.

Path prediction is useful in a wide range of applications. Most of the existing solutions, however, are based on eager learning methods where models and patterns are extracted from historical trajectories and then used for future prediction. Since such approaches are committed to a set of statistically significant models or patterns, problems can arise in dynamic environments where the underlying models change quickly or where the regions are not covered with statistically significant models or patterns. We propose a "semi-lazy" approach to path prediction that builds prediction models on the fly using dynamically selected reference trajectories. Such an approach has several advantages. First, the target trajectories to be predicted are known before the models are built, which allows us to construct models that are deemed relevant to the target trajectories. Second, unlike the lazy learning approaches, we use sophisticated learning algorithms to derive accurate prediction models with acceptable delay based on a small number of selected reference trajectories. Finally, our approach can be continuously self-correcting since we can dynamically re-construct new models if the predicted movements do not match the actual ones. Our prediction model can construct a probabilistic path whose probability of occurrence is larger than a threshold and which is furthest ahead in term of time. Users can control the confidence of the path prediction by setting a probability threshold. We conducted a comprehensive experimental study on real-world and synthetic datasets to show the effectiveness and efficiency of our approach.

sonSQL is a MySQL variant that aims to be the default database system for social network data. It uses a conceptual schema called sonSchema to translate a social network design into logical tables. This paper introduces sonSchema, shows how it can be instantiated, and illustrates social network analysis for sonSchema datasets. Experiments show such SQL-based analysis brings insight into community evolution, cluster discovery and action propagation.

Non-rotational symmetric surfaces have been widely applied in the fields of imaging, illumination, precision measurement and laser beam reshaping. Commonly used tool nose compensation method would introduce high frequency compensation components both in X and Z direction. This high frequency motion especially that of X slide is disadvantageous for the fabrication. To solve the above problem, an algorithm based on scattered data interpolation is brought out which can achieve the tool nose compensation only in Z direction. The moving least squares (MLS) method is applied for the surface fitting to calculate the tool path, tangent vector and normal vector. Then the factors influencing the accuracy of the data interpolation are discussed. Finally, a facial 3D surface was successfully fabricated and the operation time shows a high efficiency of the path optimisation algorithm. The machined results further verified the proposed tool path generation and optimisation algorithm.

In the design and manufacture of the ultra-precision machine tools, hydrostatic slides have been applied frequently, in the state-of-the-art technology. They can provide extremely smooth and accurate motion and positioning. At design stage, the carrying capacity and the stiffness of hydrostatic slide is an important value. But, the traditional calculation method of slide stiffness used in engineering is an approximate solution, the accuracy is low. In this paper, finding an accurate calculation method of slide stiffness is present, to improve modeling accuracy. Reynolds equation is the basic equation of calculating performance of hydrostatic slide, oil film pressure distribution can be obtained by solving the Reynolds equation. But Reynolds equation is a second-order partial differential equations, it is difficult to directly solve the equation using analytic methods. Therefore, a discrete model is developed to solve the problem. A method that has been widely used is finite element analysis (FEA). Matlab PDE Toolbox provides a convenient way to solving the partial differential equations based the FEA. But it can only directly solve the partial differential equations with standard form. It is discussed how to transform the Reynolds equation into elliptic partial differential equation with standard form. Then combined with the specific boundary conditions, obtain the oil film pressure distribution through Matlab PDE toolbox, and then write program to achieve high-precision calculation of carrying capacity and stiffness of the slide.

Non-rotational symmetrical micro-structured surfaces have gained more and more applications in advanced optical systems. To obtain such surfaces with low cost and high efficiency, this paper presents the key technologies for turning of such surfaces including the diamond turning system, tool path generation and tool nose compensation algorithms. The grid based tool path generation algorithm has been proposed for on-line tool path generation. The tool nose compensation is achieved by constructing cubic Hermite interpolation curve which could improve the calculating efficiency greatly. Then several micro-structured surfaces were fabricated and we focused on the analysis of micro lens array (MLA). Height of the spherical cap, diameter of the scope, and the radius of the curvature are evaluated, their maximum deviation from designed values are -2.23%, 2.27%, and 4.29% respectively. This further verified the diamond machining is an enabling method to fabricate high precision non-rotational symmetrical micro-structured surfaces.

A new design is proposed according to the requirement of a 5-axis ultra-precision machine tool for optical manufacture. This investigation was performed to see how the different configure styles influence the static performance characteristics of externally-pressurized air bearings with orifice, such as load capacity and stiffness. Based on the static performance characteristic analysis, the optimum design is done for the externally-bearing system. Finite difference method was used to obtain the numerical solution of Reynolds Equation. Load capacity was the integration of Pressure over the solution area, while the stiffness is the derivative of load capacity to gas film thickness. The load capacity of the externally-pressurized bearing grows with the increase of supply pressure and goes up with the increase of eccentricity of shaft relative to the bearing; the stiffness turns out to be tough with high supply pressure. The stiffness gets a maximum at zero eccentricity of the spindle, and gradually reduces with the increase of the eccentricity. It has proven that there exists an optimum gas-film clearance where maximum value of stiffness is obtained. There is also an optimum gas-film clearance for maximum load capacity. It should be noted that these two values is different. The restrictor throat diameter has an important influence on the stiffness and load capacity, both of which go down while larger diameter is in use. It can be concluded that there exists an optimum gas-film clearance for aerostatic bearing-rotor system. More attentions should be paid on supply pressure, gas-film thickness and orifice diameter to obtain a spindle of good performance.

Soft starter is the equipment which can control AC motor startup safely. It can not only effectively eliminate the impact on the state grid and the load and torque in the powertrain of the motor, but also improve service life of the AC motor. For these reasons, soft starter is widely used in our working which is related closely with induction motors. Above all, in order to control the AC motor starting up stably and safely, we design a novel method namely multi-slope and current-limiting fuzzy control algorithm. This technique synthesizes the respective superiority that slope voltage control method with the current limiting function and fuzzy control algorithm owned only each other. Then we evaluate the control system based on soft starter through the Matlab/Simulink and experimental tests from the hardware circuit we have been designed. The results show that the control algorithm can control AC Motor with a small current to start and the current tend to be more smoothly than each algorithm only in the whole process. We get perfectly effective controlling of the starting of motor.

In this paper, we propose a new multi-scale connected coherence tree algorithm (MCCTA) by improving the connected coherence tree algorithm (CCTA). In contrast to many multi-scale image processing algorithms, MCCTA works on multiple scales space of an image and can adaptively change the parameters to capture the coarse and fine level details. Furthermore, we design a Multi-scale Connected Coherence Tree algorithm plus Spectral graph partitioning (MCCTSGP) by combining MCCTA and Spectral graph partitioning in to a new framework. Specifically, the graph nodes are the regions produced by CCTA and the image pixels, and the weights are the affinities between nodes. Then we run a spectral graph partitioning algorithm to partition on the graph which can consider the information both from pixels and regions to improve the quality of segments for providing image segmentation. The experimental results on Berkeley image database demonstrate the accuracy of our algorithm as compared to existing popular methods.

In this paper, we propose a new radio interference detection approach. The approach which is based on Time--Frequency Ratio Of Mixtures(TIFROM) [1] employs an antenna array containing at least one directional antenna. The antenna array (i.e., one directional antenna and one Omni-antenna) acts as a mixing matrix which is invertible. Analogous to blind source separation (BSS), we use TIFROM algorithm to determine whether there is only one source in the TF domain or not, the difference is that here we look for more-than-one-cases which indicates interference on the radio frequency we are observing, while in BSS only-one-cases is sought for. Unlike other radio interference detection methods, the proposed approach only requires slight differences in the TF distributions of the considered signals. By using TF ratios of the mixed signals, it automatically determines these multi-source TF areas and estimates the strength of interference. Detailed results concerning mixtures of speech and music signals are presented and show that this approach yields very high performance.

This paper studies how to construct cluster ensembles for high dimensional data. We examine a different approach to constructing cluster ensembles. To address high dimensionality, we focus on ensemble construction methods that build on a popular dimension reduction techniques, Locally Linear Embedding (LLE). Our ensemble constructor is based on random projection in LLE subspace. We present evidence showing that ensembles generated by new algorithms perform better than those by Principal Component Analysis with subsampling (PCASS) and Random Projection simply (RP) that proposed before. Experimental results demonstrate the effectiveness of the proposed methods on several real-world data sets.

In this paper we propose a new linear feature extraction approach called Weighted Linear Embedding (WLE). WLE combines Fisher criterion with manifold learning criterion like local discriminant embedding analysis (LDE), whereas unlike LDE that only utilizes local neighbor information it uses local information and nonlocal information simultaneously. WLE is also unlike linear discriminant analysis (LDA) that treats local information and nonlocal information equally, and it uses these two kinds of information discriminatively by utilizing the Gaussian weighting. Hence, WLE is more powerful than LDA and LDE for feature extraction. Experimental results on the PolyU finger-knuckle-print database and the PolyU palmprint database indicate that our WLE algorithm outperforms principal components analysis (PCA), LDA and LDE.

The algorithm based on a sparse representation computed by L1-minimization achieved state-of-the-art performance. However, the prerequisite hypothesis in the algorithm is that the training samples from each subject can construct a linear subspace of the corresponding subject, which requires the enough images of each object to be used as training samples. The requirement of large training sets restricts its applications in face recognition. Therefore, the improved algorithm based on SRC is proposed. Firstly, the quotient image method is improved. Then the nine basis images of each subject are generated by the improved quotient image method. Lastly, the synthetic basis images are taken to act as training set to fulfill recognition task. The experimental results show that the proposed approach is feasible and effective.

Ultra-precision machine with a fast tool servo (FTS) can fabricate many kinds of optical components with complex micro-structured surfaces, achieving sub-micrometer form accuracy and nanometer surface finish without any subsequent processing. However, it is difficult to meet the ultimate processing requirements only by operators' experience due to the complicate numerical control (NC) programs and various machining parameters. To verify the NC programs, guarantee the processing quality and improve the efficiency, a simulation system is established according to the real micro-structured surface turning system. This system includes cutting force model, platform movement model, fast tool servo model, spindle movement model, vibration model and the surface topography model. Then some simulation results as the motion locus of the tool tip, three-dimensional microstructure morphology and the surface roughness are obtained. By comparing the simulated and actual results, it can be seen that this system can simulate the actual processing, predict the final machining results and has the guidance meaning for the machining of the microstructured surfaces.

A study of 322 accidents that occurred during the road transport of hazardous materials (hazmat) in China from 2000 to 2008 was carried out. The results showed an increase in the frequency of accidents from 2000 to 2007 and a decline in 2008. More than 63% of the accidents occurred in the eastern coastal areas, 25.5% in the central inland areas, and only 10.9% in the western remote areas. The most frequent types of accident were releases (84.5%), followed by gas clouds (13.0%), fires (10.2%), no substance released due to timely measures (9.9%), and explosions (5.9%). The spatial distribution, the causes and consequences of the accidents related to the population (e.g., number of people killed, injured, evacuated, or poisoned), and environment elements were analyzed. Finally, conclusions are drawn concerning the need to improve certain safety measures in the road transport of hazmat in China.

A novel method is proposed to compute the minimum distance between two 2D or 3D NURBS curves using control polygons in an efficient and robust way. The first step is to decompose both of NURBS curves into their piecewise Bzier forms. The second step is to use a two level selection process to select a subset of all possible pairs. The first level selection uses upper-lower bounding boxes of Bzier subcurves to remove pairs. The second level selection is based on the spatial relationship test between a pair of Bzier subcurves. The third step is to use a multidimensional Newton-Raphson method to compute the approximate local minimum distances of pairs of Bzier subcurves. By comparing all local minimum distances between a pair of Bzier subcurves, it is able to find the global minimum distance. The final step is to use the multidimensional Newton-Raphson method to improve the accuracy.

Water environmental risk identification (WERI) is a basic step for industrial parks in the regional water environmental risk (WER) management. It is an effective means to control risk damage in advance and help find out the regional hazard sources. It will be helpful for putting forward corresponding countermeasures to avoid water pollution accidents. Depending on the features of risk sources, control system, as well as regional risk receptors, and taking the enterprise and its surrounding area away from 1 kilometre as the evaluation unit, integrated evaluation index system (IEIS) was developed to identify the region WER. The analytic hierarchy process (AHP) was applied to determine the weight of each factor. Taixing Economic Development Zone (TEDZ) was taken as a case study for the application of the model. The results showed that 3.57%, 50% and 46.4% of the enterprises in TEDZ had high, medium and low risks, respectively. Countermeasures were provided for the decision-makers on the risk management.

The identification process of the classical Preisach model which is based on a neural network approach is presented. The fundamental idea of this approach is to identify Preisach function by training a neural network with a set of loops whose identification function is already known. The suggested identification approach has been numerically implemented and carried out for a fast tool servo system. A comparison between measured data and model predictions suggests that the mentioned identification approach can satisfactorily yield accurate results.

A saliency detection algorithm based on manifold-based similarity adaptation is proposed. The proposed algorithm is divided into three steps. First, we segment an input image into superpixels, which are represented as the nodes in a graph. Second, a new similarity measurement is used in the proposed algorithm. The weight matrix of the graph, which indicates the similarities between the nodes, uses a similarity-based method. It also captures the manifold structure of the image patches, in which the graph edges are determined in a data adaptive manner in terms of both similarity and manifold structure. Then, we use local reconstruction method as a diffusion method to obtain the saliency maps. The objective function in the proposed method is based on local reconstruction, with which estimated weights capture the manifold structure. Experiments on four bench-mark databases demonstrate the accuracy and robustness of the proposed method.