低能见度条件下车载视频的清晰化处理研究

针对低能见度天气影响交通出行的问题,提出一种新的车载视频清晰化处理算法。首先将获取图像的颜色模型转换为HSV模型,对其中的亮度分量进行直方图均衡后,再采用暗原色先验算法恢复,最后将其与原来的色度分量和饱和度分量整合后输出。实验结果表明,该算法能较大程度提高低能见度交通图像的目标辨识度,图像对比度明显增强,颜色失真较少。     

Users' psychological perception and perceived readability of wearable devices for elderly people

As the population of elderly people in society rises, the importance of technology to assist health management is growing with the demographic shift. Wearable personal computing has been the common solution proposed by related projects to fulfil the needs of elderly people. These add-on devices are attached to the users, and result in uneasiness or discomfort while wearing them in public areas. In this study, three common wearing positions – wrist, upper arm, and neck – were investigated in terms of psychological perception and perceived readability. Twenty-four Taiwanese participants in Miaoli area were asked to wear the designed prototypes for a certain time. The participants completed questionnaires and participated in an open-ended interview. Data of psychological perceptions, visibility, and readability about wearable devices were collected. The results evidenced that the personal attributes of gender, smart device user, and requirement for medical care affect the psychological perception and user attitudes. The statistical results indicated significant differences in elderly people's attitudes towards wearable devices attached to different parts of the body. The wrist was the most favourable location to attach a wearable device.     

Efficient Algorithms for Computing a Complete Visibility Region in Three-Dimensional Space

Given a set P of polygons in three-dimensional space, two points p and q are said to be visible from each other with respect to P if the line segment joining them does not intersect any polygon in P . A point p is said to be completely visible from an area source S if p is visible from every point in S . The completely visible region CV(S, P) from S with respect to P is defined as the set of all points in three-dimensional space that are completely visible from S . We present two algorithms for computing CV(S, P) for P with a total of n vertices and a convex polygonal source S with m vertices. Our first result is a divide-and-conquer algorithm which runs in O(m ² n ² α(mn)) time and space, where α(mn) is the inverse of Ackermann's function. We next give an incremental algorithm for computing CV(S,P) in O(m ² n+mn ² α(n)) time and O(mn+n ² ) space. We also prove that CV(S,P) consists of Θ(mn+n ² ) surface elements such as vertices, edges, and faces. Received November 16, 1995; revised November 11, 1996.     

Parallel Processing for Terrain Analysis in GIS: Visibility as a Case Study

The application of parallel processing to computationally intensive GIS problems has been advocated and illustrated by many researchers over the last twenty years. Despite this, GIS users have been slow to capitalize on the potential which the technology offers. Whilst todays processors are adequate for the majority of GIS uses, some applications are too processor-intensive to be deemed viable for serial machines. This is particularly true of many digital terrain modelling applications, which has been the primary focus of parallel processing in GIS to date.This paper considers the problem of parallelizing line-of-sight (LOS) calculations in determining the visibility indices of entities such as elevation vertices in a digital terrain model (DTM). This is a requirement of site selection for a particular development, especially if visibility, or more specifically, visual intrusion is likely to be a key factor in gaining planning approval. To demonstrate the simplicity and applicability of parallelizing such GIS problems, this paper presents some parallel approaches in an efficient data organization, framework using a Transputer network. Speed-up performance can be increased by a factor of twelve using a simple network of twenty Transputers. As vast quantities of spatial data become available, particularly DTMs at larger scales and denser resolution, the demands for parallel processing will inevitably increase. It is hoped that the continued experiences of todays researchers at applying parallel processing to well-defined problems will benefit the GIS users of tomorrow.     

Dynamic Polygon Visibility Ordering for Head-Slaved Viewing in Virtual Environments

This paper presents an approach to visibility called the Viewpoint Movement Space (VpMS) algorithm which supports the concept of dynamic polygon visibility orderings for head-slaved viewing in virtual environments (VE). The central idea of the approach is that the visibility, in terms of back-to-front polygon visibility ordering, does not change dramatically as the viewpoint moves. Moreover, it is possible to construct a partition of the space into cells, where for each cell the ordering is invariant. As the viewpoint moves across a cell boundary typically only a small and predictable change is made to the visibility ordering. The cost to perform this operation represents a notable reduction when compared with the cost of resolving the visibility information from the BSP tree where the classification of the viewpoint with every node plane has to be performed. The paper demonstrates how the subdivision into such cells can represent the basic source for an acceleration of the rendering process. We also discuss how the same supportive data structure can be exploited to solve other tasks in the graphics pipeline.     

可变阈值的K-Means初始中心选择方法

K-Means算法随机选择聚类中心初始点,导致聚类器性能不稳定。对此,提出基于可变阈值的初始聚类中心选择方法（VTK-Means）。该算法选择距已有初始点距离大于一个阈值的样例作为初始聚类中心,并根据满足条件的初始聚类中心个数适当调整阈值。在10个UCI数据集上的实验结果表明,该算法性能明显优于K-Means算法。     

二阶加延时模型的阶跃响应辨识方法

提出了一种带有延迟环节的二阶连续开环系统的阶跃响应辨识方法. 这种方法可以由系统的阶跃响应采样数据构造非线性方程组, 通过对方程组求解估计出系统的参数. 在不存在延迟环节的时候, 非线性方程组将简化为线性方程组. 仿真表明算法具有较高的精度.     

自适应门限的异构无线网络预切换机制

以无线局域网（WLAN）和WiMAX网络融合构成的异构网络为研究对象,研究了多模移动终端基于FMIPv6机制完成异构无线网络之间垂直切换的操作流程。针对垂直切换过程中使用固定门限值预切换机制存在的不足,提出了一种自适应门限预切换机制,并详细分析了垂直切换过程中目标网络接入时延,从而为所提出的自适应门限值预切换机制提供了理论依据。在仿真部分,扩充了NS2仿真平台上已有的功能模块,从而验证了所提出的自适应门限预切换机制的性能。     

一种改进的小波域阈值去噪算法

在D.L.Donoho和I.M Johnston提出的多分辨分析小波阈值去噪方法的基础上,提出了一种新的双变量阈值函数.采用新的阈值函数的去噪效果无论在视觉效果上,还是在信噪比增益上和最小均方意义上均优于传统的硬阈值和软阈值,克服了采用硬阈值法去噪效果不佳和软阈值法过度光滑使信号失真的缺点.通过仿真实验结果,表明该方法的有效性和优越性.     

阈值标记的分水岭彩色图像分割

目的 针对传统分水岭算法中产生的过分割问题,提出一种基于阈值标记的分水岭彩色图像分割算法。方法 该方法将分水岭算法直接应用到原始梯度图像上而不是简化之后的图像,这样做的目的是可以保护边缘信息不受损失;利用不同尺寸结构元求取彩色图像形态学梯度,解决了关于保护边缘和图像简化之间的矛盾。同时算法设计一种阈值自动选取与标记提取方法,从梯度的低频成分中提取与物体相关的局部极小值,用这些极小值构成的二值图像强制标定原始梯度图像,在修改后的梯度上进行分水岭分割。结果 在仿真实验中,利用本文算法针对不同RGB彩色图像进行分割,获得准确、连续封闭的分割边界,与其他同类方法相比,得到符合人类视觉的最小分割区域数,同时在运行效率上也有很大提高。结论 该方法可以自适应提取标记而不需要先验知识,有效解决了分水岭算法的过分割问题,相对于传统的算法,提高了分割性能,有较好的适用性和鲁棒性,可将其应用于机器视觉、生物医学以及高光谱遥感图像分割领域。     

基于可信度阈值优化的案例推理评价分类方法

为了提高案例推理(CBR) 分类器的性能, 提出一种基于可信度阈值优化的CBR 评价分类方法. 首先, 通过一种可降低时间复杂度的改进型可信度评价策略对案例重用得到的建议解的可信度进行计算; 然后, 通过遗传算法(GA) 对可信度阈值进行迭代寻优; 接着, 根据得到的优化阈值将目标案例及其建议解划分为可信集或不可信集; 最后, 对不可信集按多数重用原则进行分类结论的调整, 从而实现可信的CBR 评价分类. 对比实验表明, 改进的可信度评价策略能有效提高分类性能, 从而可提高CBR分类器的决策与学习能力.

     

标准盐度下水的受激布里渊散射阈值研究

从基本的耦合波方程出发,推导出受激布里渊散射公式,同时得出了可探测到的受激布里渊散射条件,以及间接可探测受激布里渊散射的阈值方法。采用测量激光射入不同长度样品产生SBS回波信号及透过介质激光能量强度的改变,并通过特定公式得出35‰的海水标准盐度下不同长度样品对激光的衰减情况。最后在项目所处条件下,验证了35‰海水标准盐度下受激布里渊散射阈值的存在。对受激布里渊散射阈值的产生进行了分析,并在与纯水中SBS阈值的对比中,得出了相同条件标准盐度下水的阈值较纯水中大的结果。     

提取二维图像特征点阈值的自动确定

根据灰度图像中噪声与特征点在灰度曲面中呈现出的不同分布特征,提出了对抽样区域的象素灰度进行偏差迭代运算以确定特征点象素,采用计算算术平均的办法确定阈值.采用该方法确定阈值,避免了提取图像特征点时,根据被处理图像的一些先验信息,利用试探方法确定阈值的局限性.通过对不同自然二维图像的特征点提取,证明了文中所确定的阈值是合理、有效的.     

大数据流式计算环境下的阈值调控节能策略

在大数据实时分析计算领域,流式计算的重要性不断提高,但是流式计算平台处理数据的能耗不断上升。针对这一问题,改变流式计算中节点对数据的处理方式,提出了一种阈值调控节能策略（ESTC）。首先,根据系统负载差异确定工作节点的阈值情况;其次,通过工作节点的阈值对系统数据流进行随机选择,确定不同数据处理情况调节系统的物理电压;最后,根据不同的物理电压确定系统功率。实验结果和理论分析表明,在20台普通PC机构成的流式计算集群中,实施ESTC的系统比原系统有效节能约35.2%;此外,ESTC下的性能与能耗的比值为0.0803 tuple/（s·J）,而原系统性能与能耗的比值为0.0698 tuple/（s·J）。ESTC能够在不影响系统性能的前提下,有效降低了能耗。     

Criteria for Response Monotonicity Preserving in Approximation of Fractional Order Systems

In approximation of fractional order systems, a significant objective is to preserve the important properties of the original system. The monotonicity of time/frequency responses is one of these properties whose preservation is of great importance in approximation process. Considering this importance, the issues of monotonicity preservation of the step response and monotonicity preservation of the magnitude-frequency response are independently investigated in this paper. In these investigations, some conditions on approximating filters of fractional operators are found to guarantee the preservation of step/magnitude-frequency response monotonicity in approximation process. These conditions are also simplified in some special cases. In addition, numerical simulation results are presented to show the usefulness of the obtained conditions.      

Application of a mixture model for determining the cutoff threshold for activity in high-throughput screening

Rapid throughput of assays for assessing biological activity of compounds, known as high-throughput screening (HTS), has created a need for a statistical analysis of the resulting data. Conventional methods for separating active compounds from inactive compounds based on using only a portion of the screening data to determine the cutoff threshold value are ad hoc and unsatisfactory. Taking full advantage of the entire set of screening data, we assume that the responses can be sorted into two classes; measurements associated with inactive compounds and measurements associated with active compounds. Both theoretical and practical considerations lead us to model the distribution of the measurements of inactive compounds with a Gaussian distribution. The choice is consistent with the data and our analytical experience. In our examples, active compounds inhibit enzyme and the distribution of measurements from those compounds can be characterized by a gamma or a similar one-sided long-tailed distribution. The application of this mixture of Gaussian and gamma distributions describes activity of our screening data very well. Using this model, we will show its reasoning and derivation along with describing how to optimally set the cutoff threshold value in a statistically well-defined manner. This modeling approach provides a new and useful statistical tool for HTS.     

照度场中M阵列符号图像阈值分割方法研究

针对不均匀性照度场的干扰问题,从M阵列符号源图像阈值分割出发,摒弃前期研究中惯用的固定阈值分割方法,提出了以某一点为中心的邻域亮度和阈值对应的分布关系构建成该符号源图像的阈值分布场的自适应阈值分割方法,由此能得出在该照度场中的符号图像的合理阈值.实验结果表明,新的阈值分割方法可进一步提高符号的识别效率,并为后期的三维拼接工作奠定了良好的基础.     

AS-FOA优化GRNN网络的财务预警研究

针对果蝇算法在复杂情况下寻优时易陷入局部最优等缺陷,提出一种自适应变步长果蝇算法(AS-FOA)。采用改进果蝇算法寻找GRNN网络最优参数,并利用财务数据进行危机预警以验证算法的可行性。AS-FOA算法通过赋予果蝇两次随机方向,同时引入稳定阈和适应度步长因子的概念,界定了果蝇的活跃与稳定状态,有效解决了寻优过程中因陷入局部最优而导致的收敛缓慢和低精度问题。实验表明:AS-FOA能够快速找到GRNN网络中的最佳参数,且应用于财务数据后达到的预警准确率较高。     

Behavioural threshold analysis: methodological and practical considerations for applications in information security

ABSTRACT

The application of behavioural threshold analysis to analyse group behaviour in information security presents a unique challenge in terms of the measurement instruments and methodology used to gather relevant attitude data. This paper presents an analysis of the specialised requirements for such a measurement instrument and makes methodological recommendations on the content and especially presentation of information security topics in a measurement instrument for this context. A comparison between existing methods and the specific requirements for threshold analysis is presented and serves as the main rationale for the suggested methodology. The recommended methodology and subsequent measurement instrument were implemented and experimentally tested in case studies to gauge their feasibility. Applications of behavioural threshold analysis in information security that follow the recommended methodology suggested in this article performed satisfactorily and elicits cause for further real-world experimentation.     

Long-term speech information based threshold for voice activity detection in massive microphone network

Voice activity detection (VAD) is essential for multiple microphone arrays processing, in which massive potential devices, such as microphone devices for far-field voice-based interaction in smart home environments, will be activated when sound sources appear. Therefore, the VAD can save a lot of computing resources in massive microphone arrays processing for the sparsity in sound source activity. However, it may not be feasible to obtain an accurate VAD in harsh environments, such as far-field, time-varying noise field. In this paper, the long-term speech information (LTSI) and the log-energy are modeled for deriving a more accurate VAD. First, the LTSI can be obtained by measuring the differential entropy of long-term smoothed noisy signal spectrum. Then, the LTSI is used to get labeled data for the initialization of a Gaussian mixture model (GMM), which is used to fit the log-energy distribution of noise and (noisy) speech. Finally, combining the LTSI and the GMM parameters of noise and speech distribution, this paper derives an adaptive threshold, which represents a reasonable boundary between noise and speech. Experimental results show that our VAD method has a remarkable improvement for a massive microphone network.