A cloud-based platform to develop context-aware mobile applications by domain experts

Context-awareness enables the personalization of computer systems according to the users' needs and their particular situation at a given time. The personalization capabilities are usually implemented by programmers due to the complex processes that are involved. However, an important trend in software development is that more and more software systems are being implemented not only by programmers but also by people with expertise in other domains. Since most of the existing context-aware development toolkits are designed for programmers, non-technical users cannot develop these kinds of systems. The design of tools to create context-aware systems by users that do not have programming skills but are experts in the domain where the system is going to be deployed, will contribute to speed up the adoption of these kinds of services by the society. This paper presents a cloud-based platform to ease the development of context-aware mobile applications by people without programming skills. The platform has been designed to be used in a tourism domain. This way, tourism experts can send tourist information to mobile users according to their context data (indoor/outdoor location, language, and date and time range). An energy-efficient mobile app has been developed in order to obtain context data from the user's device and to receive personalized information in real time based on these data. The architecture and implementation details of the system are presented and the evaluation of the platform by tourism domain experts is discussed.     

地理空间约束的业务流程建模方法

明晰了受地理空间约束的业务流程的两个基本概念:地理位置与地理空间信息.首先,提出了地理空间约束思想,分析了地理空间约束的内涵,并在工作流元模型的知识基础上构建了地理空间约束的UML语义框架;其次,以形式化的方式提出了地理空间约束的业务流程建模方法(LAWF-net),并描述了LAWF-net模型转化为CPN可执行模型的规则,探讨了地理空间约束对流程建模的影响;最后,通过案例结合GIS系统来扩展CPN Tools,不仅验证了LAWF-net建模方法的可行性,而且实现了地理空间约束的业务流程的可视化.     

一种基于背景自学习的高光谱图像生物信息提取方法

为了提取指纹、癌变区域等重要的生物信息,传统方法一般是使用物理、化学手段直接作用在信息载体上,这不仅需要较长时间,容易对原有信息及载体造成破坏,而且提取过程不可重现、精度较低.高光谱成像技术避免了获取信息时物理接触造成的破坏,能多次稳定获取图像,成为了一种优秀的生物信息采集途径.在此介绍一种基于背景自学习的高光谱图像信息提取方法,它解决了传统非结构化背景模型适应性不强的问题,利用空间光谱信息进一步提升了信息提取精度.实验证明,该方法能有效对背景信息进行估计,提取完整的生物目标信息,精度优于传统目标信息提取方法.     

复句关系词规则生成系统中的冲突检测与处理

复句中的关系词对研究复句中各分句的语义关系有着重要意义,在基于规则的关系词自动识别中需要大量的规则,并且规则库是动态变化和不断完善的,向规则库中入库规则时会出现规则冲突和入库错误的情况,该文探讨如何在入库时识别产生冲突的规则,并对规则进行相关的处理。对复句的普通规则、连用词规则、普通句式规则、连用句式规则四类规则进行了形式化的表示与存储,在此基础上设计了关系词检测、约束类型检测、约束条件检测、结论检测的检测流程。提出了两种冲突处理方式——优先级方式和有向无环图方式,对两种方法进行了比较。利用该检测方法和有向无环图的处理方式,入库了千余条规则。实验表明,利用该方法冲突规则的检测和处理正确率达到100%。     

使用柯西分布点扩散函数模型的单幅散焦图像深度恢复

目的 当前,大多数单幅散焦图像的3维(3D)场景深度恢复方法,通常使用高斯分布描述点扩散函数(PSF)模型,依据图像边缘散焦模糊量与场景深度的对应关系获得稀疏深度图,采用不同的扩展方法得到整个场景图像的全深度图.鉴于现有方法的深度恢复结果还不够精准,对各种噪声干扰还不够健壮,提出一种基于柯西分布的点扩散函数模型计算物体图像边缘散焦模糊量的方法.方法 将输入的单幅散焦图像分别用两个柯西分布重新模糊,利用图像边缘两次重新模糊图像间梯度比值和两个柯西分布的尺度参数,可以计算出图像中边缘处的散焦模糊量.使用matting内插方法将边缘模糊量扩展到整个图像,即可恢复场景的全深度图.结果 将原始Lenna图像旋转并加入高斯噪声以模拟图像噪声和边缘位置误差,用原图与噪声图比较了柯西分布图像梯度比值与高斯分布图像梯度比值的平均误差.使用多种真实场景图像数据,将本文方法与现有的多种单幅散焦图像深度恢复方法进行了比较.柯西分布图像梯度比值的平均误差要小于高斯分布图像梯度比值的平均误差.本文方法能够从非标定单幅散焦图像中较好地恢复场景深度,对图像噪声、不准确边缘位置和邻近边缘具有更好的抗干扰能力.结论 本文方法可以生成优于现有基于高斯模型等方法的场景深度图.同时,也证明了使用非高斯模型建模PSF的可行性和有效性.     

基于DNA序列的彩色图像加密算法

结合混沌系统和DNA密码学,提出了一种基于DNA序列的彩色图像加密算法。该算法应用了DNA序列的加法、减法、异或运算,并且把彩色图像分解为位平面进行处理。首先对彩色图像位平面分解、DNA编码;然后对DNA平面置乱、DNA加法运算、DNA异或运算;最后进行DNA解码、位平面合并,得到密文图像。实验结果表明,原始图像加密后的图像类似噪声,加密后的直方图变得更平滑,对密钥有很高的敏感性,密文图像的随机性好,密文图像相邻像素之间相关性低。     

Dialogue intent classification with character-CNN-BGRU networks

Multimedia Tools and Applications - Dialogue intent classification plays a significant role in human-computer interaction systems. In this paper, we present a hybrid convolutional neural network...     

A post-processing method for true random number generators based on hyperchaos with applications in audio-based generators

True random number generators (TRNG) are important counterparts to pseudorandom number generators (PRNG), especially for high security applications such as cryptography. They produce unpredictable, non-repeatable random sequences. However, most TRNGs require specialized hardware to extract entropy from physical phenomena and tend to be slower than PRNGs. These generators usually require post-processing algorithms to eliminate biases but in turn, reduces performance. In this paper, a new post-processing method based on hyperchaos is proposed for software-based TRNGs which not only eliminates statistical biases but also provides amplification in order to improve the performance of TRNGs. The proposed method utilizes the inherent characteristics of chaos such as hypersensitivity to input changes, diffusion, and confusion capabilities to achieve these goals. Quantized bits of a physical entropy source are used to perturb the parameters of a hyperchaotic map, which is then iterated to produce a set of random output bits. To depict the feasibility of the proposed post-processing algorithm, it is applied in designing TRNGs based on digital audio. The generators are analyzed to identify statistical defects in addition to forward and backward security. Results indicate that the proposed generators are able to produce secure true random sequences at a high throughput,which in turn reflects on the effectiveness of the proposed post-processing method.     

初等数学应用题题意理解中的常识库系统构建

类人计算领域, 题意的机器理解是数学应用题自动求解的难点. 常识性知识的缺失直接影响到题意理解的准确性. 本研究以常识为研究对象, 收集了历年初等数学古典概型的典型案例, 分析了古典概型类应用题的常识特征, 并进行了常识类型划分; 设计了XML结构存储常识性知识, 构建常识库系统实现古典概型常识的分类、表征及存储, 辅助计算机进行题意理解. 通过典型案例的应用, 其结果显示本研究构建的常识库对古典概型应用题的题意正确理解是十分有帮助的.     

CLM:differential privacy protection method for trajectory publishing

In order to solve the problem existing in differential privacy preserving publishing methods that the independent noise was easy to be filtered out,a differential privacy publishing method for trajectory data (CLM),was proposed.A correlated Laplace mechanism was presented by CLM,which let Gauss noises pass through a specific filter to produce noise whose auto-correlation function was similar with original trajectory series.Then the correlated noise was added to the original track and the perturbed track was released.The experimental results show that the proposed method can achieve higher privacy protection and guarantee better data utility compared with existing differential privacy preserving publishing methods for trajectory data.