Recently there has been an increasing interest in applying random walk based methods to recommender systems. We employ a Gaussian random field to model the top-N recommendation task as a semi-supervised learning problem, taking into account the degree of each node on the user-item bipartite graph, and induce an effective absorbing random walk (ARW) algorithm for the top-N recommendation task. Our random walk approach directly generates the top-N recommendations for individuals, rather than predicting the ratings of the recommendations. Experimental results on the two real data sets show that our random walk algorithm significantly outperforms the state-of-the-art random walk based personalized ranking algorithm as well as the popular item-based collaborative filtering method.

Developing tools for monitoring the correlations among thousands of financial data streams in an online fashion can be interesting and useful work. We aimed to find highly correlative financial data streams in local patterns. A novel distance metric function slope duration distance (SDD) is proposed, which is compatible with the characteristics of actual financial data streams. Moreover, a model monitoring correlations among local patterns (MCALP) is presented, which dramatically decreases the computational cost using an algorithm quickly online segmenting and pruning (QONSP) with O(1) time cost at each time tick t, and our proposed new grid structure. Experimental results showed that MCALP provides an improvement of several orders of magnitude in performance relative to traditional naive linear scan techniques and maintains high precision. Furthermore, the model is incremental, parallelizable, and has a quick response time.

Privacy is a critical requirement in distributed data mining. Cryptography-based secure multiparty computation is a main approach for privacy preserving. However, it shows poor performance in large scale distributed systems. Meanwhile, data perturbation techniques are comparatively efficient but are mainly used in centralized privacy-preserving data mining (PPDM). In this paper, we propose a light-weight anonymous data perturbation method for efficient privacy preserving in distributed data mining. We first define the privacy constraints for data perturbation based PPDM in a semi-honest distributed environment. Two protocols are proposed to address these constraints and protect data statistics and the randomization process against collusion attacks: the adaptive privacy-preserving summary protocol and the anonymous exchange protocol. Finally, a distributed data perturbation framework based on these protocols is proposed to realize distributed PPDM. Experiment results show that our approach achieves a high security level and is very efficient in a large scale distributed environment.

Mobility and resource-limitedness pose challenging issues to service configuration for quality of service (QoS) management in ubiquitous computing environments. Previous configuration approaches, such as static resource reservation, dynamic resource allocation and single service composition are not valid in the environments. In this study, we present an adaptive service configuration approach. Firstly, we reduce the dynamic configuration process to a control model which aims to achieve the variation of critical QoS on minimal level with less resource cost. Secondly, to deal with different QoS variations, we design two configuration strategies—service chain reconfiguration and QoS parameter adjustment—and implement them based on fuzzy logic control theory. Finally, a configuration algorithm is developed to flexibly employ the two configuration strategies in tune with the error of critical QoS in configuration process. The results of simulation experiments suggest that our approach outperforms existing configuration approaches in both QoS improvement and resource utilization.

Effective and robust recognition and tracking of objects are the key problems in visual surveillance systems. Most existing object recognition methods were designed with particular objects in mind. This study presents a general moving objects recognition method using global features of targets. Targets are extracted with an adaptive Gaussian mixture model and their silhouette images are captured and unified. A new objects silhouette database is built to provide abundant samples to train the subspace feature. This database is more convincing than the previous ones. A more effective dimension reduction method based on graph embedding is used to obtain the projection eigenvector. In our experiments, we show the effective performance of our method in addressing the moving objects recognition problem and its superiority compared with the previous methods.

Tracking multiple people under occlusion and across cameras is a challenging question for discussion. Furthermore, the cameras in this study are used to extend the field of view, which are distinguished from the same field of view. Such correspondence between multiple cameras is a burgeoning research subject in the area of computer vision. This paper effectively solves the problems of tracking multiple people who pass from one camera to another and segmenting people under occlusion using probabilistic models. The probabilistic models are composed of blob model, motion model and color model, which make the most of the space, motion and color information. First, we present a color model that uses maximum likelihood estimation based on non-parametric kernel density estimation. Second, we introduce a blob model based on mean shift, which segments the body into many regions according to the color of each person in order to spatially localize the color features corresponding to the way people are dressed. Clothes can be any mixture of colors. Third, we bring forward a motion model based on statistical probability which indicates the movement position of the same person between two successive frames in a single camera. Finally, we effectively unify the three models into a general probabilistic model and attain a maximization likelihood probability image, which is used to segment the foreground region under occlusion and to match people across multiple cameras.

This paper presents a detailed analysis of the effects of noise (reverberation) on the focusing performance of decomposition of the time reversal operator (DORT) in a noise-limited case and in a reverberation-limited case, respectively. Quantitative results obtained from simulations and experiments are presented. The results show the DORT method can be efficiently applied to target detection with enough source level to yield significant backscatter. For a target placed on the bottom, the influence of the reverberation on the focusing performance is slight. However, distinguishing between a target and constant backscattering returning from strong local clutter on the bottom (false alarms) needs further research.

We present an efficient spherical parameterization approach aimed at simultaneously reducing area and angle distortions. We generate the final spherical mapping by independently establishing two hemisphere parameterizations. The essence of the approach is to reduce spherical parameterization to a planar problem using symmetry analysis of 3D meshes. Experiments and comparisons were undertaken with various non-trivial 3D models, which revealed that our approach is efficient and robust. In particular, our method produces almost isometric parameterizations for the objects close to the sphere.

An integrated and reliable phase unwrapping algorithm is proposed based on residues and blocking-lines detection, closed contour extraction and quality map ordering for the measurement of 3D shapes by Fourier-transform profilometry (FTP). The proposed algorithm first detects the residues on the wrapped phase image, applies wavelet analysis to generate the blocking-lines that can just connect the residues of opposite polarity, then carries out the morphology operation to extract the closed contour of the shape, and finally uses the modulation intensity information and the Laplacian of Gaussian operation of the wrapped phase image as the quality map. The unwrapping process is completed from a region of high reliability to that of low reliability and the blocking-lines can prevent the phase error propagation effectively. Furthermore, by using the extracted closed contour to exclude the invalid areas from the phase unwrapping process, the algorithm becomes more efficient. The experiment shows the effectiveness of the new algorithm.

The grating eddy current displacement sensor (GECDS) for distance or position measurement used in watertight electronic calipers was described. The sensor relies on repetitive variation of inductance against displacement caused by the change of coupling areas between moving coils and static reflectors. The investigations focused on setting up and utilizing a computer model of the 3D eddy current fields and geometry to analyze causes of the production of measurement blind areas, and to investigate effects of the sensor parameters, such as axial gap between coils and reflectors, reflector length and reflector width on characteristics of the sensor. Simulation results indicated that the sensor has the smallest nonlinearity error of 0.15%, which agrees well with the experimental results.

When satellite navigation receivers are equipped with multiple antennas, they can deliver attitude information. In previous researches, carrier phase differencing measurement equations were built in the earth-centered, earth-fixed (ECEF) coordinate, and attitude angles could be obtained through the rotation matrix between the body frame (BF) and the local level frame (LLF). Different from the conventional methods, a hybrid algorithm is presented to resolve attitude parameters utilizing the single differencing (SD) carrier phase equations established in LLF. Assuming that the cycle integer ambiguity is known, the measurement equations have attitude analytical resolutions by using simultaneous single difference equations for two in-view satellites. In addition, the algorithm is capable of reducing the search integer space into countable 2D discrete points and the ambiguity function method (AFM) resolves the ambiguity function within the analytical solutions space. In the case of frequency division multiple access (FDMA) for the Russian Global Orbiting Navigation Satellite System (GLONASS), a receiver clock bias estimation is employed to evaluate its carrier phase. An evaluating variable and a weighted factor are introduced to assess the integer ambiguity initialization. By static and dynamic ground experiments, the results show that the proposed approach is effective, with enough accuracy and low computation. It can satisfy attitude determination in cases of GPS alone and combined with GLONASS.

Smart cars are promising application domain for ubiquitous computing. Context-awareness is the key feature of a smart car for safer and easier driving. Despite many industrial innovations and academic progresses have been made, we find a lack of fully context-aware smart cars. This study presents a general architecture of smart cars from the viewpoint of context-awareness. A hierarchical context model is proposed for description of the complex driving environment. A smart car prototype including software platform and hardware infrastructures is built to provide the running environment for the context model and applications. Two performance metrics were evaluated: accuracy of the context situation recognition and efficiency of the smart car. The whole response time of context situation recognition is nearly 1.4 s for one person, which is acceptable for non-time critical applications in a smart car.

A diode-triggered silicon controlled rectifier (DTSCR) is being developed as an electrostatic discharge (ESD) protection device for low voltage applications. However, DTSCR leaks high current during normal operation due to the Darlington effect of the triggering-assist diode string. In this study, two types of diode string triggered SCRs are designed for low leakage consideration; the modified diode string and composite polysilicon diode string triggered SCRs (MDTSCR & PDTSCR). Compared with the conventional DTSCR (CDTSCR), the MDTSCR has a much lower substrate leakage current with a relatively large silicon cost, and the PDTSCR has a much lower substrate leakage current with similar area and shows good leakage performance at a high temperature. Other DTSCR ESD properties are also investigated, especially regarding their layout, triggering voltage and failure current.

This study presents a new method of 4-pipelined high-performance split multiply-accumulator (MAC) architecture, which is capable of supporting multiple precisions developed for media processors. To speed up the design further, a novel partial product compression circuit based on interleaved adders and a modified hybrid partial product reduction tree (PPRT) scheme are proposed. The MAC can perform 1-way 32-bit, 4-way 16-bit signed/unsigned multiply or multiply-accumulate operations and 2-way parallel multiply add (PMADD) operations at a high frequency of 1.25 GHz under worst-case conditions and 1.67 GHz under typical-case conditions, respectively. Compared with the MAC in 32-bit microprocessor without interlocked piped stages (MIPS), the proposed design shows a great advantage in speed. Moreover, an improvement of up to 32% in throughput is achieved. The MAC design has been fabricated with Taiwan Semiconductor Manufacturing Company (TSMC) 90-nm CMOS standard cell technology and has passed a functional test.