Search

Background and Purpose: Whether imaging parameters would independently predict stroke recurrence in low-risk minor ischemic stroke (MIS) or transient ischemic attack (TIA) according to traditional score system (such as ABCD 2 score, which was termed on the basis of the initials of the five factors: age, blood pressure, clinical features, duration, diabetes) remains unclear. We sought to evaluate the association between imaging parameters and 1-year stroke recurrence in patients with TIA or MIS in different risk stratum stratified by ABCD 2 score. Methods: We included patients with TIA and MIS (National Institutes of Health Stroke Scale score ≤3) with complete baseline vessel and brain imaging data from the Third China National Stroke Registry III. Patients were categorized into different risk groups based on ABCD 2 score (low risk, 0–3; moderate risk, 4–5; and high risk, 6–7). The primary outcome was stroke recurrence within 1 year. Multivariable Cox proportional-hazards regression models were used to assess whether imaging parameters (large artery stenosis, infarction number) were independently associated with stroke recurrence. Results: Of the 7140 patients included, 584 patients experienced stroke recurrence within 1 year. According to the ABCD 2 score, large artery stenosis was associated with higher stroke recurrence in both low-risk (adjusted hazard ratio, 1.746 [95% CI, 1.200–2.540]) and moderate-risk group (adjusted hazard ratio, 1.326 [95% CI, 1.042–1.687]) but not in the high-risk group ( P >0.05). Patients with multiple acute infarctions or single acute infarction had a higher risk of recurrent stroke than those with no infarction in both low- and moderate-risk groups, but not in the high-risk group. Conclusions: Large artery stenosis and infarction number were independent predictors of 1-year stroke recurrence in low-moderate risk but not in high-risk patients with TIA or MIS stratified by ABCD 2 score. This finding emphasizes the importance of early brain and vascular imaging evaluation for risk stratification in patients with TIA or MIS.

© 1993-2012 IEEE. In this study, we present a comprehensible classifier AFSNN that embeds a new type of coherence membership function, which builds upon the theoretical findings of the axiomatic fuzzy set (AFS) theory into the hidden layer of neural network with random weights (NNRWs). Borrowing from the idea of NNRWs that employs the random initialization technique, the relation among attributes, simple concepts, and complex concepts are randomly determined. Complex concepts are generated through the combination of randomly selected simple concepts by AFS logic operation. The output weights of NNRWs are utilized to evaluate the confidence of each complex concept for every target class, which means that the feasibility of complex concepts for every class is determined analytically rather than through the tuning parameters of constraint conditions such as in conventional AFS-based classifiers. For the proposed method, compared to other neural-network-based classification methods, the fuzzy descriptions generated from complex concepts in hidden layer make classification result human understandable. We have experimented with several benchmark datasets and compared the results with other neural network-based classifiers. We show that our method outperforms Ensemble, EvRBFN, NNEP, LVQ, and iRProp+ in the seven out of ten datasets. The results show that the performance of AFSNN is competitive in terms of classification accuracy and the network shows a distinctive capability of providing explicit knowledge in the form of linguistic description.

Minority ultra-high definition video resources with ultra-high resolution will cause the occupation of large storage space, so the mainstream network rate is difficult to satisfy the requirement for smooth playback. To address the above problem, we first adopt distributed storage technology to resampling and splitting the ultra-high definition video. Then, multi-level dynamic image frame and resolution archiving are established and combined with image blocks information to form dynamic image playback archiving. This method effectively processes the conflict between ultra-high-definition video and the upload bandwidth of the server network, and thus solves a kind of ultra-high definition video broadcast problem with little consistency between frames and the need for local amplification of images. Moreover, it provides a way of broadcasting and displaying for the inheritance and protection of minority culture.

To improve the convenience of triangulation based image warping algorithm, an automatic approach for facial expression deformation is presented. In order to guarantee the expression deformation is dealt with high quality and conveniently, the approach uses two key technologies: the PSO detection and the affine transformation. The PSO algorithm is utilized to search the optimal position of rectangle containing face and eyes. The affine transformation is applied to warp related triangles to deform expression. The simulation results show that this method can produce natural smile expression, without marking the spots manually, especially works well for some shy smile expressions.

Since lacking of effective management, grid has difficulty to display its advantage. The key problem is the dynamics mechanism of resource configuration. Concerning the resource sharing strategy of grid user, the dynamics model of rational user based on the best-response dynamic mechanism is established. With typical case, the evolution trend of user sharing strategy is analyzed. The result shows that the maximize sharing of grid resources cannot be achieved because of the bounded rationality of grid user. So, the original resource management should be improved to adapt to the actual situation of grid environment.

The life evolutionary model is researched and simulated as well as other typical cellular automaton model. Different consequences are simulated with various cell numbers, different initial states and different evolution rules. After results comparison among these methods, the factors impacting stability are analyzed. Results show that the complexity of cellular automaton can provide a simple and visual tool for the theory of complexity science.

A method for precise face detection and segmentation combining with the particle swarm optimization is proposed. The method uses three key technologies: skin color segmentation, particle swarm search and curve approximation. The experimental results show that this method can eliminate the interference factor, and improve the accuracy of expression recognition, especially well for some single expression recognition.

Hexagonal phase SnS2 nanoflakes have been synthesized by reactions between an organotin precursor tetrabutyltin [TBT, (CH2CH2CH2CH3)4Sn] and carbon disulfide in hexanes at 180–200°C for 10-40 h. The structure, morphologies, composition, and properties have been characterized by powder X-ray diffraction (XRD), transmission electron microscopy (TEM), selected area electron diffraction (SAED), X-ray photoelectron spectroscopy (XPS), ICP-AES, and Raman and Mossbauer spectroscopies. XRD patterns determined the hexagonal SnS2 with lattice parameters a=3.6384 A, c=5.9201 A obtained in n-hexane, and a=3.6389 A, c=5.9288 A in cyclohexane. The flakelike morphologies were mainly caused by the anisotropic growth of SnS2. A possible mechanism is given in the paper.

Image classification is an important task in content-based image retrieval, which can be regarded as an intermediate component to handle large-scale image datasets for improving the accuracy of image retrieval. Traditional image classification methods generally utilize Support Vector Machines (SVM) as image classifier. However, there are several drawbacks of using SVM, such as the high computational cost and large number of parameters to be optimized. In this paper we propose an Extreme Learning Machine (ELM) based Multi-modality Classifier Combination Framework (MCCF) to improve the accuracy of image classification. In this framework: (i) three ELM classifiers are trained by exploring three kinds of visual features respectively, (ii) a probability based fusion method is then proposed to combine the prediction results of each ELM classifier. Experiments on the widely used TRECVID datasets demonstrate that our approach can effectively improve the accuracy of image classification and achieve performance at extremely high speed.

Based on Independent Component Analysis (ICA) and fuzzy membership functions, a new method of face image feature extraction and recognition is proposed. The new face recognition algorithm processes classification and identification of face images with combination of Independent Component Analysis and membership functions. The experiment results on ORL face databases show that the recognition algorithm is feasible and has good recognition capability.

This paper proposed a chaotic image scrambling scheme based on sorting transformation. The algorithm calculates the permuting address codes by sorting the chaotic sequence. Therefore, the scheme does not require knowing the probability density function of the chaotic orbits in advance, thus not only facilitating the choice of chaotic systems but also reducing the complexity of the scheme. The paper also presented performance analysis of the scheme by statistical methods. The results indicate that the scheme can provide high level of security owing to the strong irregularity of sorting transformation.

Web community discovery is much important for engineering optimization, search engine design and so on. In this paper, we propose a Linear Multi-Community Partition Algorithm for Scale-free network. We first take a function to measure the weight of each node in network and generate an order series by the weight of node, then iterate the algorithm to complete web community discovery. We use different scale network to test our algorithm, the experimental results indicate that this model can effectively recognize multi-communities of scale-free network and has good fitness to complex network with non- uniform density.

Ethnic Facial Feature is one of the most important face features. We create a face database of ethnic groups and extract facial features by using face recognition technology. In the feature extraction method, we adapt the algebra and geometry features from face database. In algebra features, LDA algorithm extracting the algebraic features of human face images is used. The paper also constructs a new face template to extract the geometric features and locates the points of face templates by using Gabor Wavelet. KNN and C5.0 Classifiers are used to learn the train dataset. The result indicates that the average recognition accuracy rates of Tibetan, Uighur and Zhuang ethnic groups can reach 79% by algebraic features and 90.95% by geometry features.

—Community network discovery algorithm requires many different community network data to evaluate the algorithm performance. The data sets currently we use are from the reality. During the experiment, these data sets require a large number of pre-work, but the range of their parameters is relatively limited. This paper presents two random community network generation algorithms. They can respectively generate uniform community network and non-uniform community network. At the same time, this paper builds a visible platform. Our experiment results show that the algorithms for generating random community network are feasible and can be used as test data sets generation to test community discovery algorithm.