应用射频技术的商场智能导购系统

设计的电子导购系统,包括电子标签的选择,嵌入式 Android 系统的选择,电子导购系统功能模块的组成。同时设计了客户端的两大功能模块,包括促销信息功能模块的设计,电子地图定位模块的设计。讨论了定位RFID 标签的布局,并利用 RSSI 来评估其接收的信号强度并计算出标签和读写器的距离,根据相邻的四个标签的坐标,采用极大似然估计方法获得智能购物车的位置,能够用于商场内精准的定位。     

在GPU上实现基于高斯映射的通用刀具扫描体建模

以针对数控加工仿真的需要,以高斯映射理论为基础、GPU为主要硬件平台,提出一种基于并行计算的快速通用刀具空间扫描体建模方法.首先结合通用刀具的数学模型,利用高斯映射理论分析刀具平动时的扫描体建模原理;然后分析刀具做任意空间运动时的刀具表面点法向量和对应的接触映射,推导求解扫描体包络边界的表达式;再根据GPU上的通用计算方法和GPU图形绘制管线设计基于GPU实现刀具空间扫描体建模的算法流程,并利用C++、Open GL和Open GL着色语言GLSL实现了该方法.通过刀具扫描体建模实例,验证了文中方法的正确性与实时性.     

自上而下注意图分割的细粒度图像分类

目的 针对细粒度图像分类中的背景干扰问题,提出一种利用自上而下注意图分割的分类模型。方法 首先,利用卷积神经网络对细粒度图像库进行初分类,得到基本网络模型。再对网络模型进行可视化分析,发现仅有部分图像区域对目标类别有贡献,利用学习好的基本网络计算图像像素对相关类别的空间支持度,生成自上而下注意图,检测图像中的关键区域。再用注意图初始化GraphCut算法,分割出关键的目标区域,从而提高图像的判别性。最后,对分割图像提取CNN特征实现细粒度分类。结果 该模型仅使用图像的类别标注信息,在公开的细粒度图像库Cars196和Aircrafts100上进行实验验证,最后得到的平均分类正确率分别为86.74%和84.70%。这一结果表明,在GoogLeNet模型基础上引入注意信息能够进一步提高细粒度图像分类的正确率。结论 基于自上而下注意图的语义分割策略,提高了细粒度图像的分类性能。由于不需要目标窗口和部位的标注信息,所以该模型具有通用性和鲁棒性,适用于显著性目标检测、前景分割和细粒度图像分类应用。     

基于SAM-CS和SOFM的胃上皮肿瘤细胞图像识别

针对胃上皮肿瘤细胞图像(以下简称肿瘤细胞图像)黏结严重和信息冗余的特点,提出了一种将自适应观测矩阵的压缩感知(SAM-CS)和自组织特征映射(SOFM)神经网络相结合的算法。该算法将肿瘤细胞图像拉成列向量,然后利用通过自适应过程产生的观测矩阵,基于压缩感知理论对图像信息进行观测,产生线性观测向量,最后利用SOFM神经网络的学习算法对观测向量进行训练和分类,实现对肿瘤细胞图像的识别。实验表明,相比常用算法,该算法至少提高了4.2%的识别准确率和5.7%的运算速度。     

A dimension map for molecular aggregates

A pair of gyradius ratios, defined from the principal radii of gyration, are used to generate a dimension map that describes the geometry of molecular aggregates in water and in organic solvents. Molecular dynamics simulations were performed on the aggregation of representative biomolecules and polyaromatic compounds to demonstrate application of the dimension map. It was shown that molecular aggregate data on the dimension map were bounded by two boundary curves, and that the map could be separated into three regions representing three groups of structures: one-dimensional rod-like structures; two-dimensional planar structures or short-cylinder-like structures; and three-dimensional sphere-like structures. Examining the location of the aggregates on the dimension map and how the location changes with solvent type and solute material parameter provides a simple yet effective way to infer the aggregation manner and to study solubility and mechanism of aggregation.     

Optimization of neural network model using modified bat-inspired algorithm

The success of an artificial neural network (ANN) strongly depends on the variety of the connection weights and the network structure. Among many methods used in the literature to accurately select the network weights or structure in isolate; a few researchers have attempted to select both the weights and structure of ANN automatically by using metaheuristic algorithms. This paper proposes modified bat algorithm with a new solution representation for both optimizing the weights and structure of ANNs. The algorithm, which is based on the echolocation behaviour of bats, combines the advantages of population-based and local search algorithms. In this work, ability of the basic bat algorithm and some modified versions which are based on the consideration of the personal best solution in the velocity adjustment, the mean of personal best and global best solutions through velocity adjustment and the employment of three chaotic maps are investigated. These modifications are aimed to improve the exploration and exploitation capability of bat algorithm. Different versions of the proposed bat algorithm are incorporated to handle the selection of the structure as well as weights and biases of the ANN during the training process. We then use the Taguchi method to tune the parameters of the algorithm that demonstrates the best ability compared to the other versions. Six classifications and two time series benchmark datasets are used to test the performance of the proposed approach in terms of classification and prediction accuracy. Statistical tests demonstrate that the proposed method generates some of the best results in comparison with the latest methods in the literature. Finally, our best method is applied to a real-world problem, namely to predict the future values of rainfall data and the results show satisfactory of the method.     

Efficient and secure data hiding in encrypted images: A new approach using chaos

ABSTRACT

The recent trends in technology have made it possible to reproduce and share digital media more easily and more rapidly. This leads to the problem of exploiting the media illegitimately. To alleviate this problem, many cryptographic techniques are used to secure the data by encrypting them. However, the cloaked form of encrypted data attracts the intruder and shows the way to malicious attacks. Due to this, in recent times steganography has drawn more attention to secure the data. This article presents a new technique that embeds data in intermediate significant bit (ISB) and least significant bit (LSB) planes of the cover image. The method employs chaotic maps to generate random locations to hide the data bits as well as permutation order to encrypt the cover image. The cover image is first encrypted by applying permutation order, then embedding is carried out on the random locations generated. After embedding, the decrypted cover image is transmitted. This method provides two-level security in extracting the hidden data. Experimental outcomes (PSNR, MSE, NAE, and NCC) confirm that the method is proficient. The randomness of the values generated by chaotic maps is assessed by the NIST standard test suite.     

鄱阳湖湖区及五河七口DEM制作

阐述了在ArcGis平台上利用数字地形图生产鄱阳湖湖区及五河七口DEM制作的方法.     

RFID与影像地图技术在光缆线路维护管理中的应用

通过对光缆线路维护管理现状的分析,探讨了实现光缆线路维护管理信息化的技术和方法,对其中的SOA架构、RIA技术和地理信息系统的应用做了详细的论述．同时也对光缆线路维护管理信息系统设计与实现中的RFID设备信息自动读取方法、影像地图技术实现做了具体的描述。     

基于用户查询日志的命名实体挖掘

针对大规模查询日志中丰富的命名实体的挖掘是数据挖掘领域中的重要研究课题。已有的研究工作提出了一种基于种子实体的抽取框架,利用实体间的分布相似度进行挖掘。然而该工作只有当种子实体仅属于单个语义类别时才能取得好的结果,实际上命名实体往往可能从属于多个类别。该文通过引入一个弱指导话题模型,利用少量的人工指导信息,很好地解决了实体的类别模糊性,提高了挖掘的有效性。实验表明该文提出的方法在实体挖掘性能上显著优于已有的方法。