人脸识别技术在广播电视节目内容监管系统中的应用与实现

为提升广播电视节目内容监测监管效率,节约人工成本,实现自动发现和标记节目中出现的敏感人物,本文将人脸识别技术引入到广播电视智慧监管系统建设中。文章论述了人脸识别引擎的工作流程,以敏感人物样本库数据为识别样本,引用人脸检测算法和人脸检索算法,对视频节目内容抽帧后的图像进行敏感人物识别,标识并存储人物信息。其中,对人脸检测算法和人脸检索算法的原理和实现流程进行了探讨研究。

Weed Classification Using Particle Swarm Optimization and Deep Learning Models

Weed is a plant that grows along with nearly allfield crops,including rice,wheat,cotton,millets and sugar cane,affecting crop yield and quality.Classification and accurate identification of all types of weeds is a challenging task for farmers in earlier stage of crop growth because of similarity.To address this issue,an efficient weed classification model is proposed with the Deep Convolutional Neural Network(CNN)that implements automatic feature extraction and performs complex feature learning for image classification.Throughout this work,weed images were trained using the proposed CNN model with evolutionary computing approach to classify the weeds based on the two publicly available weed datasets.The Tamil Nadu Agricultural University(TNAU)dataset used as afirst dataset that consists of 40 classes of weed images and the other dataset is from Indian Council of Agriculture Research–Directorate of Weed Research(ICAR-DWR)which contains 50 classes of weed images.An effective Particle Swarm Optimization(PSO)technique is applied in the proposed CNN to automa-tically evolve and improve its classification accuracy.The proposed model was evaluated and compared with pre-trained transfer learning models such as GoogLeNet,AlexNet,Residual neural Network(ResNet)and Visual Geometry Group Network(VGGNet)for weed classification.This work shows that the performance of the PSO assisted proposed CNN model is significantly improved the success rate by 98.58%for TNAU and 97.79%for ICAR-DWR weed datasets.

基于FPGA集群的Office口令恢复优化实现

口令恢复是口令找回和电子取证的关键技术,而加密的Office文档被广泛使用,实现Office加密文档的有效恢复对信息安全具有重要的意义。口令恢复是计算密集型任务,需要硬件加速来实现恢复过程,传统的CPU和GPU受限于处理器结构,大大限制了口令验证速度的进一步提升。基于此,文中提出了基于FPGA集群的口令恢复系统。通过详细分析Office加密机制,给出了各版本Office的口令恢复流程。其次,在FPGA上以流水线结构优化了核心Hash算法,以LUT(Look Up Table)合并运算优化改进了AES(Advanced Encryption Standard)算法,以高速并行实现了口令生成算法。同时,以多算子并行设计了FPGA整体架构,实现了Office口令的快速恢复。最后,采用FPGA加速卡搭建集群,配合动态口令切分策略,充分发掘了FPGA低功耗高性能的计算特性。实验结果表明,无论在计算速度还是能效比上,优化后的FPGA加速卡都是GPU的2倍以上,具有明显的优势,非常适合大规模部署于云端,以缩短恢复时间找回口令。

A Method of Clustering Components into Modules Based on Products' Functional and Structural Analysis

Modularity is the key to improving the cost-variety trade-off in product development. To achieve the functional independency and structural independency of modules, a method of clustering components to identify modules based on functional and structural analysis was presented. Two stages were included in the method. In the first stage the products’ function was analyzed to determine the primary level of modules. Then the objective function for modules identifying was formulated to achieve functional independency of modules. Finally the genetic algorithm was used to solve the combinatorial optimization problem in modules identifying to form the primary modules of products. In the second stage the cohesion degree of modules and the coupling degree between modules were analyzed. Based on this structural analysis the modular scheme was refined according to the thinking of structural independency. A case study on the gear reducer was conducted to illustrate the validity of the presented method.

数值流形法中确定不规则单元的多边形求交法

由数学覆盖和物理网格形成的流形单元的积分区域可以有任意形状,其形状为物理网格和数学网格的相交部分,所以流形单元积分区域的形成为两个多边形(数学网格和物理网格)的求交运算。文章提出了以节点操作求解两多边形交集的方法和算法,对于实心任意多边形的任意相交方式,该方法是准确和通用的,且易于计算机程序的实现。

城市突发事件中基于事故演变的救援需求决策模型及其优化求解

针对城市突发事件中待救点对应急救援物资的需求状态随着事故的演变而发生变化的情况,将待救点对应急救援物质的需求设计成马尔可夫决策过程,并提出一种动态的物资调配策略,构建救援物资需求决策模型,然后通过花朵授粉算法进行优化求解。某市突发地震应急物资调配实例证明,本文方法能够给出合理的救援需求满足策略,使得待救点对物资的需求更低,需求变化更加平稳。花朵授粉算法对模型求解的效果相比粒子群算法和萤火虫算法有更快的收敛速度和更优的求解结果。

基于改进粒子群算法的煤矿井下相机参数优化

煤矿井下视觉测量精度的关键在于相机参数的准确性,提出一种改进粒子群算法对张氏标定法所得参数进行优化,在迭代初期和后期分别使用e指数和粒子适应度函数值动态调节惯性权重。经过实验,优化后相机参数平均误差与标准差相比张氏标定法分别提高了23.65%、22.83%。且改进PSO算法与传统PSO算法、e指数调节惯性权重PSO相比,可以明显降低迭代次数,提高优化精度,是提高煤矿井下视觉测量准确性的一种有效方法。

A network security situation prediction model based on wavelet neural network with optimized parameters

The security incidents ion networks are sudden and uncertain, it is very hard to precisely predict the network security situation by traditional methods. In order to improve the prediction accuracy of the network security situation, we build a network security situation prediction model based on Wavelet Neural Network （WNN） with optimized parameters by the Improved Niche Genetic Algorithm （INGA）. The proposed model adopts WNN which has strong nonlinear ability and fault-tolerance performance. Also, the parameters for WNN are optimized through the adaptive genetic algorithm （GA） so that WNN searches more effectively. Considering the problem that the adaptive GA converges slowly and easily turns to the premature problem, we introduce a novel niche technology with a dynamic fuzzy clustering and elimination mechanism to solve the premature convergence of the GA. Our final simulation results show that the proposed INGA-WNN prediction model is more reliable and effective, and it achieves faster convergence-speed and higher prediction accuracy than the Genetic Algorithm-Wavelet Neural Network （GA-WNN）. Genetic Algorithm-Back Propagation Neural Network （GA-BPNN） and WNN.

Bernoulli-based random undersampling schemes for 2D seismic data regularization

Seismic data regularization is an important preprocessing step in seismic signal processing. Traditional seismic acquisition methods follow the Shannon–Nyquist sampling theorem, whereas compressive sensing（CS） provides a fundamentally new paradigm to overcome limitations in data acquisition. Besides the sparse representation of seismic signal in some transform domain and the 1-norm reconstruction algorithm, the seismic data regularization quality of CS-based techniques strongly depends on random undersampling schemes. For 2D seismic data, discrete uniform-based methods have been investigated, where some seismic traces are randomly sampled with an equal probability. However, in theory and practice, some seismic traces with different probability are required to be sampled for satisfying the assumptions in CS. Therefore, designing new undersampling schemes is imperative. We propose a Bernoulli-based random undersampling scheme and its jittered version to determine the regular traces that are randomly sampled with different probability, while both schemes comply with the Bernoulli process distribution. We performed experiments using the Fourier and curvelet transforms and the spectral projected gradient reconstruction algorithm for 1-norm（SPGL1）, and ten different random seeds. According to the signal-to-noise ratio（SNR） between the original and reconstructed seismic data, the detailed experimental results from 2D numerical and physical simulation data show that the proposed novel schemes perform overall better than the discrete uniform schemes.

A Semi-Supervised WLAN Indoor Localization Method Based on l1-Graph Algorithm

For indoor location estimation based on received signal strength( RSS) in wireless local area networks( WLAN),in order to reduce the influence of noise on the positioning accuracy,a large number of RSS should be collected in offline phase. Therefore,collecting training data with positioning information is time consuming which becomes the bottleneck of WLAN indoor localization. In this paper,the traditional semisupervised learning method based on k-NN and ε-NN graph for reducing collection workload of offline phase are analyzed,and the result shows that the k-NN or ε-NN graph are sensitive to data noise,which limit the performance of semi-supervised learning WLAN indoor localization system. Aiming at the above problem,it proposes a l1-graph-algorithm-based semi-supervised learning( LG-SSL) indoor localization method in which the graph is built by l1-norm algorithm. In our system,it firstly labels the unlabeled data using LG-SSL and labeled data to build the Radio Map in offline training phase,and then uses LG-SSL to estimate user's location in online phase. Extensive experimental results show that,benefit from the robustness to noise and sparsity ofl1-graph,LG-SSL exhibits superior performance by effectively reducing the collection workload in offline phase and improving localization accuracy in online phase.

Simple Insulin Dose Adjustment Using 3-3-1 Algorithm in Japanese Patients with Type 2 Diabetes: Start Kanazawa Study (Self-Titration Aggressive Algorithm with Glargine Trial)

We implemented a 3-3-1 algorithm in order to provide safe and simple self-titration in patients who newly initiated BOT as well as who were already on BOT and evaluated its utility in clinical setting. A total of 46 patients, 21 patients in the newly-initiated group and 25 patients in the existing BOT group performed dose adjustment using 3-3-1 algorithm. HbA1c was significantly improved 4 weeks after the initiation from 8.5% ± 1.2% at baseline to 7.3% ± 0.7% at the final evaluation (p 0.01, vs. Baseline). The average daily insulin units increased throughout the study period from 10.1 ± 6.7 at baseline to 14.6 ± 8.9 units at the final evaluation. Weight didn’t significantly change throughout the study (p = 0.12). The incidents of hypoglycemia were 0.8/month during the insulin dose self-adjustment period and 0.4/month during the follow-up period. The 3-3-1 algorithm using insulin glargine provided a safe and simple dose adjustment and demonstrated its utility in patients who were newly introduced to insulin treatment as well as who were already on BOT.

ACCURACY OF THE RETRIEVED TEMPERATURE AND HUMIDITY FIELDS FOR TYPHOON HAIYAN UTILIZING THE ADVANCED TECHNOLOGY MICROWAVE SOUNDER

One-dimensional retrieval was performed on Typhoon Haiyan utilizing the advanced technology microwave sounder onboard the satellite Suomi NPP to retrieve the temperature and water vapor profiles of the typhoon.Comparisons of the retrieved profiles and ECMWF reanalysis were made to assess the results. The main conclusions are as follows.(1) The results have high spatial resolution and therefore can precisely represent the temperature and humidity distribution of the typhoon.(2) The retrieved temperature is low in the areas of low temperature and high in the areas of high temperature; similar patterns are observed for humidity. This means that systematic revision may be needed during routine application.(3) The results of the retrieved temperature and humidity profiles are generally accurate, which is quite important for typhoon monitoring.

Study on correlativity among capacity dimension D_0,informationd imension D_1,algorithmic complexityC

ＳｔｕｄｙｏｎｃｏｒｒｅｌａｔｉｖｉｔｙａｍｏｎｇｃａｐａｃｉｔｙｄｉｍｅｎｓｉｏｎＤ;，ｉｎｆｏｒ－ｍａｔｉｏｎｄｉｍｅｎｓｉｏｎＤ;，ａｌｇｏｒｉｔｈｍｉｃｃｏｍｐｌｅｘｉｔｙＣ（ｎ）ａｎｄｂＶａｌｕｅＷＥＩ－ＢＩＮＨＡＮ（韩渭宾），ＧＵＩ－Ｘ...

方家山核电厂发电机氢气泄漏率计算方法改进

氢冷发电机大都运行在手动补氢的方式,发电机氢气泄漏率的计算也采用人工提取数据,手动输入软件计算的方式。该方法在方家山核电厂的实际应用中存在计算结果波动幅度较大,难以及时发现泄漏率上升的问题。另外,由于氢冷系统采用手动方式补氢,氢气的压力呈现出锯齿型,无法用通用的算法计算出连续的泄漏率。本研究定量分析了泄漏率计算波动大的原因,给出了提高精度降低波动幅度的建议,并首次在业界提出了一种克服锯齿型压力波动,计算实时泄漏率的算法。对比结果和实际应用表明,实时泄漏率的计算结果准确度更高,采用连续的趋势跟踪更容易发现泄漏率的上升。新的计算方法在业界具用推广的价值。

Experimental implementation of a fixed-point duality quantum search algorithm in the nuclear magnetic resonance quantum system

In this work, we demonstrated a fixed-point quantum search algorithm in the nuclear magnetic resonance (NMR) system. We constructed the pulse sequences for the pivotal operations in the quantum search protocol. The experimental results agree well with the theoretical predictions. The generalization of the scheme to the arbitrary number of qubits has also been given.

An integrated model for structure optimization and technology screening of urban wastewater systems

The conventional approach to wastewater system design and planning considers each component separately and does not provide the optimum performance of the entire system. However, the growing concern for environmental protection, economic efficiency, and sus- tainability of urban wastewater systems requires an integrated modeling of subsystems and a synthetic evaluation of multiple objectives. In this study, a multi- objective optimization model of an integrated urban wastewater system was developed. The model encom- passes subsystems, such as a sewer system, stormwater management, municipal wastewater treatment, and a wastewater reclamation system. The non-dominated sort- ing genetic algorithm （NSGA-II） was used to generate a range of system design possibilities to optimize conflicting environmental and economic objectives. Information from a knowledge base, which included rules for generating treatment trains as well as the performance characteristics of commonly used water pollution control measures, was utilized. The trade-off relationships between the objec- tives, total water pollution loads to the environment, and life cycle costs （which consist of investment as well as operation and maintenance costs）, can be illustrated using Pareto charts. The developed model can be used to assist decision makers in the preliminary planning of system structure. A benchmark city was constructed to illustrate the methods of multi-objective controls, highlight cost- effective water pollution control measures, and identify the main pressures on urban water environment.

ITERATIVE l1 MINIMIZATION FOR NON-CONVEX COMPRESSED SENSING

An algorithmic framework, based on the difference of convex functions algorithm （D- CA）, is proposed for minimizing a class of concave sparse metrics for compressed sensing problems. The resulting algorithm iterates a sequence ofl1 minimization problems. An exact sparse recovery theory is established to show that the proposed framework always improves on the basis pursuit （l1 minimization） and inherits robustness from it. Numerical examples on success rates of sparse solution recovery illustrate further that, unlike most existing non-convex compressed sensing solvers in the literature, our method always out- performs basis pursuit, no matter how ill-conditioned the measurement matrix is. Moreover, the iterative l1 （ILl） algorithm lead by a wide margin the state-of-the-art algorithms on l1/2 and logarithimic minimizations in the strongly coherent （highly ill-conditioned） regime, despite the same objective functions. Last but not least, in the application of magnetic resonance imaging （MRI）, IL1 algorithm easily recovers the phantom image with just 7 line projections.

Prediction of Protein Homo-Oligomer Types by Using IB1 Algorithm

Protein homo-oligomers play an important role in various vital activities. Successful prediction of protein homo-oligomers directly from primary sequence is very beneficial to understand their protein function. In this paper, a total of 14 238 homo-oligomeric protein sequences are predicted by IB1 algorithm. 10-fold cross-validation test is applied to test the predictive capability of the proposed method. The predictive results show that overall prediction accuracy is 90.46%, which is at least 9% higher than that of previous results; furthermore,the sensitivity and Matthew＇s correlation coefficient for each class of homo-oligomers are also improved significantly. The results show that IB1 algorithm is effective and feasible,and very suitable for predicting protein homo-oligomer types.

Learning robust principal components from L1-norm maximization

Principal component analysis(PCA) is fundamental in many pattern recognition applications.Much research has been performed to minimize the reconstruction error in L1-norm based reconstruction error minimization(L1-PCA-REM) since conventional L2-norm based PCA(L2-PCA) is sensitive to outliers.Recently,the variance maximization formulation of PCA with L1-norm(L1-PCA-VM) has been proposed,where new greedy and nongreedy solutions are developed.Armed with the gradient ascent perspective for optimization,we show that the L1-PCA-VM formulation is problematic in learning principal components and that only a greedy solution can achieve robustness motivation,which are verified by experiments on synthetic and real-world datasets.

The holographic reconstructing algorithm and its error analysis about phase-shifting phase measurement

Phase-shifting measurement and its error estimation method were studied according to the holographic principle.A function of synchronous superposition of object complex amplitude reconstructed from N-step phase-shifting through one integral period(N-step phase-shifting function for short)was proposed.In N-step phase-shifting measurement,the interferograms are seen as a series of in-line holo-grams and the reference beam is an ideal parallel-plane wave.So the N-step phase-shifting function can be obtained by multiplying the interferogram by the original reference wave.In ideal conditions,the proposed method is a kind of synchro-nous superposition algorithm in which the complex ampli-tude is separated,measured and superposed.When error exists in measurement,the result of the N-step phase-shifting function is the optimal expected value of the least-squares fitting method.In the above method,the N+1-step phase-shifting function can be obtained from the N-step phase-shifting function.It shows that the N-step phase-shifting function can be separated into two parts:the ideal N-step phase-shifting function and its errors.The phase-shifting errors in N-steps phase-shifting phase measurement can be treated the same as the relative errors of amplitude and intensity under the understanding of the N+1-step phase-shifting function.The difficulties of the error estimation in phase-shifting phase measurement were restricted by this error esti-mation method.Meanwhile,the maximum error estimation method of phase-shifting phase measurement and its formula were proposed.