Metabolic P Systems： A Discrete Model for Biological Dynamics

A recent methodology to model biochem- ical systems is here presented. It is based on a concep- tual framework rooted in membrane computing and de- veloped with concepts typical of discrete dynamical sys- tems. According to our approach, from data observed at suitable macroscopic temporal scales, one can deduce, by means of algebraic and algorithmic procedures, a dis- crete model （called Metabolic P system） which accounts for the experimental data, and opens the possibility to under- stand the systemic logic of the investigated phenomenon. The procedures of such a method have been implemented within a computational platform, a Java software called MetaPlab, processing data and simulating behaviors of metabolic models. In the paper, we briefly describe the theory underlying the modeling of biochemical systems by Metabolic P systems, along with its development stages and the related extensive literature.     

Automatic Recognition Algorithm of AM Signals Based on Spectrum and Modulation Characters

To meet the actual requirement of automatic monitoring of the shortwave signals under wide band ranges, a technique for automatic recognition is studied in this paper. And basing upon the spectrum and modulation characters of amplitude modulation （AM） signals, an automatic recognition scheme for AM signals is proposed. The proposed scheme is achieved by a joint judgment with four different characteristic parameters. Experiment results indicate that the proposed scheme can effectively recognize AM signals in practice.     

A Multilevel Security Model for Private Cloud

Towards data leak caused by misoperation and malicious inside users, we proposed a multilevel se- curity model based on Bell-lapadula （BLP） model. In our model each subject was assigned with a security level. Sub- jects can read objects only when their security levels are not less than objects＇ security levels, and subjects can write objects only when their security levels are not more than objects＇ security levels. The current security level in our model can be dynamically changed when users read sensi- tive data, since users can access data with different security levels in private cloud. Our model use mandatory access control method to control user＇s operation and can guar- antee that users can not leak sensitive data after they read them. Our model can be proved secure by mathematical method, and we implemented a prototype system of our model and the experimental results show that it is secure.     

Automatic Image Annotation Using Semi-Supervised Multi-Instance Multi-Label Learning Algorithm

Automatic image annotation has emerged as an important research topic. From the perspective of machine learning, the annotation task fits both multiinstance and multi-label learning framework due to the fact that an image is composed of multiple regions, and is associated with multiple keywords as well. In this paper, we propose a novel Semi-supervised multi-instance multi-label （SSMIML） learning framework, which aims at taking full advantage of both labeled and unlabeled data to address the annotation problem. Specifically, a reinforced diverse density algorithm is applied firstly to select the Instance prototypes （IPs） with respect to a given keyword from both positive and unlabeled bags. Then, the selected IPs are modeled using the Gaussian mixture model （GMM） in order to reflect the semantic class density distribution. Furthermore, based on the class distribution for a keyword, both positive and unlabeled bags are redefined using a novel feature mapping strategy. Thus, each bag can be represented by one fixed-length feature vector so that the manifold-ranking algorithm can be used subsequently to propagate the corresponding label from positive bags to unlabeled bags directly. Experiments on the Corel data set show that the proposed method outperforms most existing image annotation algorithms.     

Unified Maneuvering Model for Target Tracking by Using Range-rate Measurement

A novel Unified maneuvering model （UMM） for tracking the target performed the complex curvilinear motion in two dimensions is presented, where the model is constructed by time-varying parameters and driven by the along-track and cross-track acceleration in- put, and the range rate plays a key role to update vari- able parameters online. In Cartesian coordinates, by incorporating the cross-track acceleration component which is estimated online by using the range-rate, and the along- track acceleration component which is determined under the assumption of zero-mean first order Markovian process, the proposed UMM exhibits highly self-adjustment capability to compensate the mismatch between the actual motion and the mathematic model adaptively, espe-cially, shows a well capability to approximate the standard Interacting multiple model （IMM） under the situation of complex curvilinear motion and a higher level of the measurement noise. Simulation results validate our theory and show that UMM-based filtering is superior to that using the Current statistic model （CSM）, and has a well approx- imation to IMM, meanwhile, has a low computational load substantially.     

Microblog Burst Keywords Detection Based on Social Trust and Dynamics Model

Microblog has emerged as a popular medium for providing new sources of information and rapid communications, particularly during burst topics. Burst keywords detection from real-time microblog streams is important for burst topics detection. The exiting algorithms may detect fake burst keywords without taking into account the trustworthiness of the users and human＇s daily timetable. Our work is the first to combine the trustworthiness of the users with burst keywords detection. We propose a novel approach to detect burst keywords based on social trust and dynamics model. We adapt basic notions of dynamics from physics and model keywords bursts as momentum change of the keywords. On the analogy of physical dynamics model, this approach defines mass as the trustworthiness of user and position as the frequency of keywords. We compute each keyword＇s burst value by using Moving average convergence/divergence （MACD） and determine whether it is a burst keyword in a given time window. The experimental results on large-scale Sina microblog dataset show that the proposed approach can avoid detecting fake burst keywords.     

Viewpoint Optimization Using Genetic Algorithm for Flying Robot Inspection of Electricity Transmission Tower Equipment

To fully exploit the limited flight-time of the flying robot, and ensure the successful visibility of target, viewpoint optimization is proposed in this paper for the inspection of electricity transmission tower equipment with an optimization function to determine the best view- points in a local viewpoint region. The local viewpoint regions are generated from the local objective regions which are determined by the geometrical structure of a priori 3D model for the electricity transmission tower equipment. The optimization function is structured based on three factors including visibility, viewing quality and observation distance. In addition, the fitness function of genetic algo-rithm is used to find the optimal viewpoint. The experimental results demonstrate the effectiveness and efficiency of the proposed viewpoint selection algorithm.     

An Efficient Causal Structure Learning Algorithm Based on Recursive Simultaneous Equations Model

A new algorithm, the RSEM （Recursive simultaneous equations model） algorithm, is presented for causal structure learning under the LSEM （Linear structural equations model）. The algorithm effectively applies recursive simultaneous equations model to causal structure learning. This paper makes two specific contributions. Firstly, under the assumption that knowing the causal order of the variables, we show that recursive simultaneous equations model can be used for causal structure learning under the LSEM regardless of whether the datasets follow multivariate Gaussian distribution. Secondly, the performance of the RSEM algorithm is compared with the state-of-the-art algorithms on 7 networks. Simulation results show that the RSEM algorithm outperforms existing algorithms in terms of time performance, and has a quite high accuracy for thresholds 0.005 and 0.01.     

PCNN Model Analysis and Its Automatic Parameters Determination in Image Segmentation and Edge Detection

The Pulse coupled neural network （PCNN） has been widely used in digital image processing, but the automatic parameters determination is still a difficult as- pect, which becomes the focus of PENN research. In this paper, by the classical solution to difference equations and the time-domain analysis of PCNN model, we provide the expressions of the firing time and the firing period of neu- rons, and reveal the “mathematics firing” phenomenon of PCNN. Based on this, we propose a new method of auto- matic parameters determination based on both eliminating the “mathematics firing” and getting the highest efficiency of PCNN. We also present an edge detection model on the basis of image segmentation of PCNN and a method to determine automatically the parameters of the model. Ex- perimental results prove the validity and efficiency of our proposed algorithm for the segmentation and the edge de- tection of the test images.     

An Improved Cochlear Implant Strategy Incorporating Frequency Modulation Information

Some currently popular strategies in Cochlear implants （CIs） fail to encode the temporal fine structure cues, which are crucial for speech perception in noise or melody appreciation of CI patients. We propose an improved strategy based on the CIS （Continuous interleaved sampling） model by introducing partial lowfrequency temporal fine structure cues or Frequency modulation （FM） information into the slowly varying temporal envelops. A psychoacoustic experiment was conducted to validate the improved strategy by measuring the Mandarin vowel, tone, consonant and sentence recognitions on normal hearing listeners. Experimental data show that the introduction of frequency modulation information can improve the CI performance greatly, especially for vowel and tone perception. Firstly even at the most severe noise condition the vowel perception can get nearly half intelligibility, and the tone recognition scores increased over 20% at various noise conditions. Secondly under moderate noise conditions or in quiet the fine structure cues also contribute significantly to the consonant and sentence recognitions. Finally, the proposed strategy has its own application values because it does not introduce too many high-frequency components into the model, which can not be perceived by deaf patients. .     

A Novel Refined Track Initiation Algorithm for Group Targets Based on Group Model

Traditional refined track initiation methods for group targets have mistakes or loss of tracks when tracking irregular motions, for the reason that they rely on a stable relative position of group members. To solve the problem, a group dynamic model was introduced for proposing a new initiation algorithm and its whole framework. We made a self-adaptive improvement of the group separation on various group radii. After the pre-association of these groups, a state equation derived from the model was used for predictions of group members. Then a relational matrix was defined for refined data associations. Finally tracks were validated by logic-based method. Particular scenarios and Monte Carlo simulations showed that, compared with algorithms based on relative position, this algorithm has better performance on the adaptability to changes of a group structure and the correctness of initiation.     

Short Texts Classification Through Reference Document Expansion

With the rapid development of information technology, short texts arising from socialized human inter- action are gradually predominant in network information streams. Accelerating demands are requiring the industry to provide more effective classification of the brief texts. However, faced with short text documents, each of which contains only a few words, traditional document classifi- cation models run into difficulty. Aggressive documents expansion works remarkably well for many cases but suf- fers from the assumption of independent, identically dis- tributed observations. We formalize a view of classification using Bayesian decision theory, treat each short text as ob- servations from a probabilistic model, called a statistical language model, and encode classification preferences with a loss function defined by the language models and the ex- ternal reference document. According to Vapnik＇s meth- ods of Structural risk minimization （SRM）, the optimal classification action is the one that minimizes the struc- tural risk, which provides a result that allows one to trade off errors on the training sample against improved gener- alization performance. We conduct experiments by using several corpora of microblog-like data, and analyze the ex- perimental results. With respect to established baselines, results of these experiments show that applying our pro- posed document expansion method produces better chance to achieve the improved classification performance.     

A Power Control Scheme for Target Localization in WSN

In this paper, we focus on the target tracking problem in sensor networks and propose an Power-saving target localization scheme （PSTL） based on a conjec- ture model that reflects the moving pattern ofa target, and also a corresponding two-step communication protocol be- tween Base station （BS） and sensors. BS executes a query mechanism to determine which sensors should be used for detailed information according to a limited amount of data received from sensors. This scheme reduces both energy consumption and communication bandwidth requirement, prolongs the lifetime of the wireless sensor networks. Simulation results indicate that it can achieve a high accuracy while saving a large amount of energy.     

＂Smart Cafe＂： A Mobile Local Computing System Based On Indoor Virtual Cloud

With network developing and virtualization rising, more and more indoor environment （POIs） such as care, library, office, even bus and subway can provide plenty of bandwidth and computing resources. Meanwhile many people daily spending much time in them are still suffering from the mobile device with limited resources. This situation implies a novel local cloud computing paradigm in which mobile device can leverage nearby resources to facilitate task execution. In this paper, we implement a mobile local computing system based on indoor virtual cloud. This system mainly contains three key components： 1）As to application, we create a parser to generate the ＂method call and cost tree＂ and analyze it to identify resource- intensive methods. 2） As to mobile device, we design a self-learning execution controller to make offtoading decision at runtime. 3） As to cloud, we construct a social scheduling based application-isolation virtual cloud model. The evaluation results demonstrate that our system is effective and efficient by evaluating CPU- intensive calculation application, Memory- intensive image translation application and I/ O-intensive image downloading application.     

A Novel Coupled Bi-directional Diffused Image Inpainting Model with Morphological Invariant Property

In this paper, a novel coupled inpainting model which bi-directionally diffuses image information is proposed. It is morphological invariant which restores the target region based on geometric property. The image information is diffused along both the direction normal to edges and along isophotes. Total variation （TV） model is used to diffuse along the direction normal to edges which reconnects the broken lines in the target region directly. The along isophotes diffused part is the inviscid Helmholtz vorticity equation in fluid dynamics. The Helmholtz equation diffuses a smooth measure of image along isophotes, and it is morphological invariant. The novel model is comprised by 2nd order Partial differential equations （PDEs）, so the numerical scheme of it is simple and the processing time is limited. Experimental results demonstrate that the novel model smoothly restores the target region by diffusing along two orthogonal directions and preserves the linear structure which leads to its better performance than conventional inpainting models.     

Design and Implementation of a Driver＇s Eye State Recognition Algorithm Based on PERCLOS

In order to improve the general detection accuracy of eye state, this paper puts forward an innovative method for judging human eye state based on PERCLOS. After pretreatment of the eye image, Hough transformation is used for ellipse detection and pupil position. The gray projection variance threshold analysis is then used to help make the final detection. Freeman chain and the Snake model algorithm are used for the corner detection and pre- cise calculation of the height of an open eye. Thus the PER- CLOS value and the eye state can be figured out. The performance of our eye state recognition algorithm is validated by more than 1000 images within product database. The statistics result shows that the fatigue detection accuracy rate can meet the need of usage in complex environment.     

A Novel Method for Analyzing the Noise Characteristic of Solar Cells with En-In Model

In this paper, a novel method for analyzing the noise characteristic of solar cells with En-In model was studied. The En-In noise model in two-port network was introduced to study the low-frequency noise characteristic of solar cells. According to the relationship between the output noise power spectrum and the two noise parameters in En-In noise model of the solar cell, known as En and In, an accurate method for extracting the two noise parame- ters was proposed. At the same time, the measurement method for the both parameters from 1Hz to 10kHz was studied. After 1/f noise curve fitting and characteristic frequency of G-R noise extraction on the noise spectrums of a large amount of solar ceils, the analyzing results of spectrum compositions proved the validity and significance of the En-In noise model for noise analysis of solar cells. It also provided the essential theoretical and experimental basis for the further research on noise characteristic and reliability estimation of solar cells and PV modules.     

On-line/Off-line Threshold Proxy Re-Signatures

To improve the performance of threshold proxy re-signatures, the notion of on-line/off-line threshold proxy re-signatures is introduced. The bulk of re-signature computation can be done in the off-line phase before the message arrives. The results of this pre-computation are saved and then utilized in the on-line phase once a message to be re-signed is known. Based on any threshold proxy re-signature scheme and a threshold version of chameleon hash function, we present a generic on-line/off-line thresh- old proxy re-signature scheme that can convert any ex- isting secure threshold proxy re-signature scheme into an on-line/off-line one. The on-line phase of our scheme is ef- ficient： computing a re-signature share requires one round of communication, two modular additions and one mod- ular multiplication. Our scheme is provably secure under the discrete logarithm assumption without random oracles. Moreover, our scheme can achieve robustness in the pres- ence of [n/3] malicious proxies.     

Virtual Network Mapping Through Locality-aware Topological Potential and Influence Node Ranking

Efficiently mapping multiple independent Virtual networks （VNs） over a common infrastructure sub- strate is a challenging problem on cloud computing plat- forms and large-scale future Internet testbeds. Inspired by the idea of data fields, we apply a topological poten- tial function to node ranking and propose an algorithm called Locality-aware node topological potential ranking （LNTPR）, which can precisely and efficiently reflect the relative importance of nodes. Using LNTPR and the con- cept of locality awareness, we develop the Locality-aware influence choosing node （LICN） algorithm based on a node influence model that considers the mutual influence be- tween a mapped node and its candidate mapping nodes. LNTPR and LICN improve the integration of node and link mapping. Simulation results demonstrate that the proposed algorithms exhibit good performance in deter- mining revenue, acceptance ratio, and revenue/cost ratio.