Abstract: A clustering algorithm named "Clustering by fast search and find of density peaks" is for finding the centers of clusters quickly. Its accuracy excessively depended on the threshold, and no efficient way was given to select its suitable value, i.e., the value was suggested be estimated on the basis of empirical experience. A new way is proposed to automatically extract the optimal value of threshold by using the potential entropy of data field from the original dataset. For any dataset to be clustered, the threshold can be calculated from the dataset objectively instead of empirical estimation. The results of comparative experiments have shown the algorithm with the threshold from data field can get better clustering results than with the threshold from empirical experience.
Abstract: Query on uncertain data has received much attention in recent years, especially with the development of Location-based services (LBS). Little research is focused on reverse k nearest neighbor queries on uncertain data. We study the Probabilistic reverse k nearest neighbor (PRkNN) queries on uncertain data. It is succinctly shown that, PRkNN query retrieves all the points that have higher probabilities than a given threshold value to be the Reverse k-nearest neighbor (RkNN) of query data Q. The previous works on this topicmostly process with k > 1. Some algorithms allow the cases for k > 1, but the efficiency is inefficient especially for large k. We propose an efficient pruning algorithm-Spatial pruning heuristic with louer and upper bound (SPHLU) for solving the PRkNN queries for k > 1. The experimental results demonstrate that our algorithm is even more efficient than the existent algorithms especial for a large value of k.
Abstract: In most of traditional P systems, each rule has the same execution time. That way of using the rules is not quite realistic from a biological point of view, because external conditions always change in an unpredicted manner such that different reaction may take different time to execute. In this work, we investigate the computation efficiency of tissue P systems by removing the restriction that each rule should complete in one time unit. The timed tissue P system is constructed by adding a time mapping to the rules to specify the execution time for each rule. A uniform and time-free solution to 3-coloring problem is proposed, where the execution time of the computational processes involved can vary arbitrarily and the output produced is always the same.
Abstract: A low-cost low-power area-efficient receiver front-end prototype for X-band applications is presented. The front-end primarily consists of 3 blocks: a singleended input differential-output Low noise amplifier (LNA), a double balanced down-converter, and Inter-frequency (IF) buffers providing single-ended output. Including are also Low dropout regulators (LDO) and Electrostatic discharge (ESD) protection circuits coinciding with the forgoing blocks. Local oscillator (LO) frequency is chosen such that output signal locates in L/S-band for extending subsequent applications. Experimentally exhibiting a conversion gain of around 44dB with 7.5-dB Single-sideband (SSB) Noise figure (NF), the front-end totally draws 24mA from an external 3.3-V supply. Fabricated in a 65-nm CMOS technology, this compact receiver occupies an area of only 0.22mm2.
Abstract: Data sparseness brings significant challenges to the research of recommender systems. It becomes more severe for neighborhood-based collaborative filtering. We introduce the trust relation computing of the sociology field. Instead of the traditional similarity computing method, the trust degree is integrated for the nearest neighbor selection. The trust network is constructed by the expansion of different path length, and the trust value between the users can be obtained by the trust transmission rules. To verify the effectiveness of our method, we give the experiments on different techniques for rating prediction, including Pearson based method, the User position similarity (UPS) based method and the trust with Pearson and UPS. We also give the t-test result. The implementation of the experiment on the Epinions data set shows that the proposed method can improve the system performance significantly.
Abstract: A low-power high-linearity Analog frontend (AFE) composed by a Digital-to-analog converter (DAC) and a Low-pass filter (LPF) is proposed in this paper for ZigBee transmitter applications. The DAC is realized by current-steering topology which adopts an optimized segmentation method to resolve the contradiction between requirements of linearity and power. A Successive approximation register (SAR) frequency auto-tuning operation is presented for the LPF to accommodate the performance deterioration due to the Process, voltage and temperature (PVT) variations. Implemented in a 0.13μm CMOS technology, the proposed AFE has been fully integrated in a ZigBee transceiver chip with an area of 0.23mm2. The experimental results demonstrate that it achieves a linearity of 30dBm Output 3rd order intercept point (OIP3) and dissipates 4.75mA from a 1.2V supply.
Abstract: The task of real-time microblog filtering is to decide if the subsequently posted tweets are relevant to a given query representing special information needs. The filters based on the retrieval model or the text classification model are the main solutions for this task. To best exploit the strengths of the two models, a hybrid model using the retrieval model as prior knowledge to rectify the hyperplane of classification is proposed. The hybrid filtering model incorporates the language model and the logistic regression model. Evaluated on the Text RetriEval Conference (TREC) 2012 microblog real-time filtering track dataset, the experimental results show that the proposed model is significantly better than the logistic regression model and the language model. Especially, it outperforms the best method of the TREC 2012 microblog real-time filtering track.
Abstract: It is a challenging task to carry on the research concerning the lateral stabilization of Single wheel robot (SWR). The lateral dynamics of earlier flywheel stabilizing SWR is derived to explain the shortcomings of such a mechanism, and the recovery torque proportional to the acceleration and moment of inertia of the flywheel is proved. But as the motor being a speed server system, it is difficult to control its acceleration to actuate the flywheel to provide recovery torque for SWR; and the moment of inertia of the flywheel must be large enough to produce adequate recovery torque, which makes such a system rather cumbersome. We proposed a new mechanism to solve the problem by introducing an electromagnetic force. The proposed mechanism is described briefly, and the dynamic analyses of such a new SWR is given, then the stability analysis is also done. Three-dimensional (3D) multi-body simulation, numerical simulation and physical invented pendulum prototype experiments were conducted to verify the proposed mechanism. The experiment results verified that the proposed mechanism is feasible and have some advantages over the flywheel balanced SWR.
Abstract: PMGI/ZEP520A/PMGI/ZEP520A fourlayer resist stack is firstly proposed for T-gates fabrication of InP-based High electron mobility transistors (HEMTs). Gate-head and gate-foot are exposed in single-step Electron beam lithography (EBL), which avoids alignment deviation by automatic self-alignment. The newly introduced PMGI at the bottom greatly improves the adhesiveness of ZEP520A resist with the substrate. The optimal gate-foot length reaches 101nm for a design of 50nm gate footprint pattern, and which can be improved to be 66.8nm for 30nm gate footprint pattern. Finally, Tgates in nanometer regime have been successfully incorporated into InP-based HEMTs fabrication. Benefiting from both the narrow gate-foot and the reduced parasitic gatecapacitance by single-step EBL technique with the fourlayer resist stack, the fabricated devices with gate-foot length of 101nm demonstrate excellent DC and RF performances: the maximum extrinsic transconductance, the current-gain cutoff frequency and maximum oscillation frequency are 1051mS/mm, 249GHz and 415GHz, respectively.
Abstract: To deal with the issues of energy consumption for Network on chip (NoC), this paper proposes a method for partitioning an NoC architecture into multiple voltage-frequency islands to achieve low energy consumption. The method considers the number of Voltagefrequency islands (VFIs), delay and reliability as multiconstraints. A mathematical model for Integer linear program (ILP) algorithm is constructed and LPSolve is used to solve the issues of VFI partition. The validity of the VFI partition method is accurately verified for E3S and MMS. Simulation results show that the proposed method is more reasonable and smaller consumption can be a chieved while multi-constraints are met. This method can get less energy consumption from 33.6% to 9.1% or 16.7% than those by the other methods.
Abstract: Ontology-based semantic retrieval can improve the efficiency of information retrieval. This paper proposes a semantic retrieval model based on domain ontology of orchard disease and pests. According to Forestry Thesaurus, we semi-automatically construct a domain ontology and repair ontology inconsistency to ensure the accuracy and uniqueness of the domain knowledge. A concept similarity algorithm is proposed and applied to calculate sentence similarity. We present a synthetic sentence similarity algorithm, which is a combination of the traditional sentence similarity algorithm and the weighted sentence similarity algorithm. Compared with other related methods through experiments, our retrieval model has higher accuracy in semantic retrieval.
Abstract: Influential user evaluation is great importance in many application areas of online social networks. In order to identify influential users in a more adequate and practical way, we propose a Dynamic regional interaction model (DRI) to evaluate user influence in online social networks. Influential users can be identified by the influence effect on different distance users based on dynamic regional interaction model. We have applied the influential user identification method to Sina Weibo and the experimental results show that compared with the existing methods the proposed method can identify the influence users in a more accuracy and efficiency way.
Abstract: Face diagnosis is one of the four diagnostic methods in Traditional chinese medicine (TCM). The morbidity of the organs can be revealed from the facial complexion. Due to the ambiguity of face diagnosis in TCM, Fuzzy support vector machine (FSVM) is utilized to remove the influence of outliers. The facial complexion recognition in TCM based on FSVM is proposed in this paper, which includes the following steps: 1) The facial cheek region are segmented as skin blocks; 2) The color feature in Lab color space is extracted from skin block to represent facial complexion characteristic; 3) The fuzzy membership is calculated to obtain the membership of the training samples, in which the final membership of training samples is determined by combining membership calculation based on distance with the facial complexion recognition based on color modeling; 4) The FSVM is built to classify facial complexion characteristic. Experimental results show that proposed facial complexion recognition has better antiinterference performance as well as a higher recognition rate up to 82%.
Abstract: Current methods of image protection based on chaos encryption can only provide security in Human visual system (HVS) level, however they can not provide usage control when the image is opened, printed or exported. To solve this problem, we proposed a novel image digital rights management scheme for Confidential image data security based on encryption and watermark (CIDSEW) with high-level security, usage control and traceability, in which we employed full content image encryption for confidentiality of the image to be protected, and we firstly proposed strict and detailed Usage control (UC) scheme for Confidential image data (CIData) usage password-based authentication, opening times, printing and exporting control. And when the CIData need to delivery or export to other users or domain, we proposed the secure export and misused tracing and detect approach, in which before the ciphered CIData is exported, we decrypted the CIData in a plain mode, and simultaneously we embedded user-identity-related and hardware-related information as robust watermark for traceability and responsibility confirmation. Finally, we evaluated the proposed CIDSEW by groups of variant size image data for security and efficiency, a large amount of groups of experiments manifest the proposed CIDSEW is secure, efficient, pervasive and robust for confident image data protection, usage control and misuse tracing.
Abstract: A novel sensor deployment method utilize Discrete wavelet transform (DWT) is proposed, and the DWT is used to calculate the sub-band energy entropy to ascertain the coverage cavities in Wireless sensor networks (WSNs). We address the problem of deploying a limit number of sensors to optimize the coverage ratio in 3D surface, while it is a complex surface in space and sensors can be deployed only onto it. Another novel aspect of this paper is that the method followed utilizes an Artificial bee colony algorithm with dynamic search strategy (ABC-DSS), which mimics the behavior of bees, and the new modified ABC-DSS algorithm matches the sensor deployment problems on 3D surface well. The extensive simulations illustrate that comparing with the deployment method based on Particle swarm optimization (PSO) and ABC, the ABC-DSS which utilizes the wavelet sub-band energy entropy is functional and efficient on 3D surface deployment problems.
Abstract: An ensemble method based on supervised learning for segmenting the retinal vessels in color fundus images is proposed on the basis of previous work of Zhu et al. For each pixel, a 36 dimensional feature vector is extracted, including local features, morphological transformation with multi-scale and multi-orientation, and divergence of vector field which is firstly used to extract feature of retinal image pixels. Then the feature vector is used as input data set to train the weak classifiers by the Classification and regression tree (CART). Finally, an AdaBoost classifier is constructed by iteratively training for the retinal vessels segmentation. The experimental results on the public Digital retinal images for vessel extraction (DRIVE) database demonstrate that the proposed method is efficient and robust on the fundus images with lesions when compared with the other methods. Meanwhile, the proposed method also exhibits high robustness on a new Retinal images for screening (RIS) database. The average accuracy, sensitivity, and specificity of improved method are 0.9535, 0.8319 and 0.9607, respectively. It has potential applications for computer-aided diagnosis and disease screening.
Abstract: In this paper, a multi-channel post-filtering approach in reverberant environment based on detection and estimation scheme is presented. A modified Signal presence probability (SPP), which is in consideration of reverberation, is proposed with a novel estimator Direct-toreverberate ratio (DRR) to adapt to distant-talking scene. SPP is known a key estimator to instruct the updating of transient noise or residual directional interference and form gain function in the time-frequency domain, consequently a new desired signal detection scheme is proposed to improve its accuracy. Appropriate spectral enhancement technique is applied to the noisy speech signal taking advantage of the modified SPP estimator. The proposed multi-channel post-filtering is tested in different nonstationary noisy and reverberant environments. Experimental results show that it achieves considerable improvement on signal preservation of the desired speech with more noise reduction over the comparative algorithms.
Abstract: Although Optical character recognition (OCR) technology has achieved huge progress in recent years, character misrecognition is inevitable. In order to realize high fidelity content of document digitalization, we propose a new Convolutional neural networks (CNN) based confidence estimationmethod.We detect the misrecognized characters through comparing the confidence value with a preset threshold, so as to leave the recognition errors as embedded images in the output digital documents. We adopted sofmax as the estimation of posteriori probability, overlap pooling and maxout with dropout technologies in CNN architecture design. Experimental results show that our method has achieved an explicit improvement compared to baseline system.
Abstract: Magnitude information and direction information are utilized to improve the spatiotemporal Kriging method in the step of variogram construction. Basement of spatiotemporal variogram construction is to calculate vector distance between time slices, and space slices. Criterions of vector distance to spatiotemporal variogram construction is discussed, and a new vector distance model considering both magnitude and direction information is proposed. Case study using data from National Network of Geomagnetic Observatories of China is carried out for cross validation. Results illustrate that the proposed model performs better in analyzing in the statistics of Mean absolute error (MAE) and Mean square error (MSE). The accuracy of spatiotemporal Kriging interpolation is improved with such an operation in vector distance.
Abstract: We investigated a simple Adaptive selection/maximal-ratio (ASM) combining cooperative system with output threshold over independent nonidentical composite Nakagami-lognormal fading channels using Mixture gamma (MG) distribution. Some novel closed-form expressions for the probability density function, the cumulative distribution function and the momentgeneration function of the output Signal-to-noise ratio (SNR) for the ASM system are derived, respectively. The average symbol error rate, outage probability and the diversity order for the ASM system are given based on the above expressions. For the purpose of comparison, we also derive the statistical characterizations of the conventional selection combining and maximal-ratio combining cooperative systems using MG distribution. Numerical and simulation results are shown to verify the accuracy of the analytical results under different scenarios, such as varying average SNR, fading parameters per hop, and the location of relaying nodes. These results show that ASM is a simple and flexible cooperative system, and can be useful in future practical deployment.
Abstract: This paper addresses the joint Feedback bit allocation (FBA) and Power allocation (PA) problem for a limited feedback coordinated multi-cell transmission system over composite fading channels. The Rayleigh/Lognormal fading channel is considered, and the effects of path loss and the macro-level correlation are also taken into account. A modified performance measure, referred to as the virtual sum rate is adopted to facilitate the analysis. The closed forms of the lower bound on virtual sum rate are derived over uncorrelated and correlated shadowed fading channels. Based on the lower bound, we then formulate an optimization problem of maximizing the virtual sum rate under constraints on the total number of feedback bits per MS and the total transmit power per BS. Considering the tight coupling between the FBA and the PA, we solve the optimization problem by solving an alternating sequence of FBA and PA subproblems using a coordinated ascent searching method. Numerical results show that the proposed strategy can converge in a few iterations and generally yield better performance than the conventional FBA schemes under composite fading channels.
Abstract: The objective of optimizing a projection matrix is to decrease the mutual coherence between a projection matrix and a basis matrix. In this paper, a novel block-based method is proposed to design a projection matrix in compressed sensing. Here, the projection matrix is divided into two blocks. The relationship between the two blocks was obtained by reasoning and proving. Theoretical analysis demonstrates that the mutual coherence between the whole projection matrix and the whole basis matrix keeps as good as the mutual coherence between the block matrix and blocked basis matrix. Experimental results show that the proposed method obtains better performance compared to existing methods.
Abstract: The conditional h-vertex connectivity of G is defined as the minimum cardinality |S| (S ⊂V (G)) such that G-S is disconnected and has minimum degree at least h. It is an important measure of fault tolerance of networks. In this paper, we prove the lower bound of conditional hvertex connectivity of any n-dimensional hypercube-like network. We also determine the conditional h-vertex connectivity of Crossed cubes, Locally twisted cubes, Möbius cubes which are the members of hypercube-like networks.
Abstract: Recent years have witnessed a rapid growth in using Internet of things (IoT), which facilitates and simultaneously raises challenges for the intelligent logistics. The real-world devices can provide their functionality as Web services. Because of the dynamicity and heterogeneity of the target networking environment, the services offered by IoT resources cannot be composed by simply extending existing Service oriented architecture (SOA) approaches. Logistics, which consists of a complex network of organizations and business processes, must be monitored in real time. Based on the proper understanding of the possible exceptions, we propose a middleware approach to solve the logistics service composition in IoT, where a decentralized coordination mechanism is used to monitor the component services with few resources efficiently. Through a set of experiments, the effectiveness and robustness of our approach are evaluated.
Abstract: Canonical Artificial bee colony (ABC) algorithm with a single species is insufficient to extend the diversity of solutions and may be trapped into the local optimal solution. This paper proposes a new co-evolutionary ABC algorithm (HABC) based on Hierarchical communication model (HCM). HCM combines advantages of global and local communication pattern. With adjustment strategies on species and groups, HCM can reduce the computational complexity dynamically. Performance tests show that the HABC algorithm exhibit good performance on accuracy, robustness and convergence speed. Compared with ABC and Integrated co-evolution algorithm (IABC), HABC performs better in solving complex multimodal functions.
Abstract: Based on the principle of phase velocity tapering or stepping output circuit, using the phase velocity variation of the output circuit, four kinds of nonuniform output circuit models have been presented in broadband helix Traveling-wave tubes (TWTs). Taking an example of 6-18GHz, 100W broadband helix TWT and using the improved one-dimensional large signal program, second harmonic suppression performances have been studied by calculating the quantitative relationship between fundamental efficiency, second harmonic output power, and the stepping (tapering) position and strength of phase velocity variation for the four kinds of nonuniform output circuits. From the research results of backward wave and second harmonic suppression feature in two kinds of nonuniform output circuits, we obtained the design principle of phase velocity stepping or tapering, which could suppress the Backward wave oscillation (BWO) and reduce the output power of second harmonic for broadband high-power helix TWTs.
Abstract: The Dynamic programming track before detect (DP-TBD) algorithm has been widely used for detection and tracking of weak targets. The selection of the merit function has an immediate influence on the performance of the DP-TBD. The amplitude merit function is easy to calculate, but the performance of which will decrease in the presence of non-Gaussian clutter. The likelihood ratio merit function in closed analytical form is difficult to derive under non-Gaussian background without target signal parameters. To solve this problem, a novel DPTBD algorithm based on local linearization is proposed. Taking maximum of the state conditional probability ratio of the target as the optimal criteria, a recursive integration equation is derived. The equation is locally linearized by Taylor series expansion and a suboptimal multi-frame test statistic is developed. The calculation of new merit function in the statistic needs only clutter distribution model, and heavy clutter peak can be restrained by making use of clutter distribution characters. So the proposed algorithm can efficiently extract weak target in strong non-Gaussian clutter. Numerical simulations are provided to assess and compare the performance of the proposed algorithm. It turns out that the proposed algorithm has better detection and tracking performance than the widely used DPTBD algorithm at present and is resilient to various clutter distribution models.
Abstract: High isolation and compact size diplexer with two cascaded triplet sections is presented in this paper. By using the quarter-wavelength resonators and stepped impedance resonator to form cascaded triplets, size reduction is achieved. By exploiting different coupling between the resonators, multiple transmission zeros can be obtained to improved the isolation between the two channels. An experimental diplexer with lower and upper channels centered at 1.8GHz and 2.2GHz is designed to verify the proposed theory. Without extra matching network, the proposed diplexer shows high isolation of greater than 41dB.