基于KIM算法和Adaboost级联的快速人脸检测

KIM算法可以有效解决帧差法易产生空洞以及背景减法不能检测出背景灰度接近目标的问题,得到准确的运动区域,再利用基于级联分类器的Adaboost算法对运动区域进行人脸检测。实验结果表明,该方法可以更加快速、准确地实现人脸检测,具有较好的实时性。     

基于ViBe算法及Canny边缘检测的运动目标提取

为了能快速提取视频监控中的运动目标,提出将ViBe算法与Canny边缘检测算法相结合快速准确提取运动目标的方法。该方法利用ViBe算法检测运动目标,运用Canny算子提取运动目标边缘轮廓,并将其与通过区域填充得到的精确运动目标相"与",获取运动目标准确的外边界;通过膨胀、腐蚀、闭运算等数学形态学图像处理,得到运动目标完整检测区域,结合相对应视频序列帧可以提取运动目标。实验表明,该方法可以快速、实时地提取多运动目标。     

新型基于帧间差分法的运动人脸检测算法研究

研究运动人脸检测、改善运动人脸检测的准确性问题,针对目前在运动人脸检测算法中存在的鲁棒性与精确性不强、适应性弱以及检测效率不高的问题.为了实时精确识别从脸图像,提出了一种帧间差分法的运动人脸检测算法.算法通过运动目标检测,获取运动人体的大致区域,利用人脸检测算法对运动区域进行检测,得到较为准确的人脸检测结果.经过对算法的仿真证明,算法能够快速、准确地检测出图像序列中的单个运动人脸,具有高效、准确的特点,有很大的实用价值.     

基于帧间差分和背景差分结合的Camshift目标跟踪算法

Camshift算法需要手动标定目标区域,且具有无法适应目标的高速运动、相似颜色背景和遮挡等情况的局限性。针对这些情况提出结合帧间差分法和背景差分相结合的方法对Camshift算法进行改进。首先利于帧间差分和背景差分相结合检测出运动目标区域。然后用该区域初始化跟踪目标窗口。当有相似颜色背景干扰或遮挡情况发生时,利用检测出的运动目标区域对搜索窗口进行限制。同时,使用Kalman滤波对下一帧的搜索窗口进行预测,从而使该算法适合高速运动目标的跟踪。实验表明该算法能够准确对目标窗口进行初始化,且在目标高速运动、遮挡、和相似颜色背景干扰情况下,仍能进行适时实时有效跟踪。     

视频序列中运动目标检测算法

提出一种结合区域级和像素级背景差分法的目标检测算法,可以有效解决视频序列中噪声分布不均问题。利用一种基于熵能的局部自适应阈值划分出前景和背景,在此基础上对前景和背景邻接区域像素点进行检测,并提出一种自适应光线变化的背景更新算法。实验结果表明,该算法比传统单阈值背景差分法抗噪能力更强,检测目标轮廓更加完整,能够准确检测出运动目标。     

基于水平集分割的3DOGHM检测算法

针对帧间差分检测运动区域抗噪性差、某些部位无法完全恢复、所提取的运动目标容易产生空洞的问题,提出一种基于水平集分割的3DOGHM运动目标检测算法,在3DOGHM分离运动区域及背景的基础上,采用一种改进的水平集进化模型进行运动目标分割。实验结果表明,该算法抗干扰能力强,可以更准确、完整地检测出运动目标。     

基于视觉显著性的均值漂移跟踪算法

为解决突变运动下的目标跟踪问题,提出了一种基于视觉显著性的均值漂移跟踪算法,将视觉注意机制运用到均值漂移跟踪框架中,利用时空显著性算法对视频序列进行检测,生成视觉显著图,从视觉显著图对应的显著性区域中建立目标的颜色特征表示模型来实现运动目标跟踪.实验结果表明:该算法在摄像机摇晃等动态场景下可以较准确检测出时空均显著的目标,有效克服了在运动目标发生丢失和遮挡等情况下跟踪不稳定的问题,具有较强的鲁棒性,从而实现复杂场景下目标较准确的跟踪.     

基于背景差分的人体运动检测

本文对视频序列中的运动目标检测方法进行了研究,提出了一种基于背景差分的运动目标检测算法.首先通过建立混合高斯模型,获取背景图像,并对背景进行实时更新.然后通过背景差分法确定运动目标区域,再进行图像处理,得到较为精确的运动目标.最后利用八连通域算法,对多个目标进行分割.实验证明,该方法能够快速、准确的检测出多个运动目标.     

自适应权值的MRF分割与跟踪方法

提出一种基于自适应权值的区域马尔可夫随机场（MRF）分割与跟踪的方法,该方法利用了相邻像素区域的空间相关性,自适应更新系统能量函数中的参数β,可以更准确地分割出运动目标,在此基础上建立分片积分直方图特征模板,并结合Kalman预测与目标运动方向等信息,进一步提高算法准确性,实现目标匹配跟踪。实验结果表明,本文算法在部分遮挡、光线变化等情况下,能准确实现运动目标分割与跟踪。     

视频中运动目标检测算法研究及实现

背景建模是视频监控中检测运动目标的一种常用方法。针对固定场景的视频,给出了一种视频运动目标检测算法。利用混合高斯建模的背景差分法,在复杂背景条件下提取出运动目标,在OpenCV库函数的支持下进行编程测试,算法可以准确地检测运动目标。实验结果表明该算法的有效性、优越性和可行性。     

Teaching mineral prospecting by computer-assisted simulation techniques

The simulation game, which involves the interaction of a geological model, a set of rules, and teaching materials, is used to teach students and professional people attending refresher courses the techniques of mineral exploration. Using geochemical and geophysical measurements, including detailed borehole logs, the students plan an exploration strategy, control the exploration budget, discover the structure and geology of the area represented by the model, outline probable mineralizations, and evaluate the ore reserves. They choose the borehole locations, the type of drilling machine, and the drilling depth, taking continually into consideration the financial aspects.The model is multidisciplinary, involving stratigraphy, sedimentology, structural geology, tectonics, climatic conditions, metallogeny, mineralogy, and petrology. Certain of these characteristics are expressed as complex mathematical functions, which are solved by the program. The computer program also prints out the detailed borehole logs, controls the budget and drilling equipment, and enforces the rules of the game.     

The Design and Role of Synthetic Creative Traits in Artificial Ant Colonies

Previous and current research on ant colonies have focused on the emergence of complex and sophisticated group-behaviors that are characteristic of the whole colony, starting from simple rules characterizing each individual ant. Examples of emergent behaviors include foraging and nest building. While social scientists, biologists, and physicists have focused on observing and analyzing the behavior of actual ants, researchers in computer science have investigated through modeling and simulation of synthetic ants on the computer, the emergent behavior in a collection of simple robots, simulated evolution of computer programs, and the formation of pheromone patterns during foraging under unlimited food supply. This paper has two objectives. First, it presents a systematic study, through behavior modeling and simulation, of the influence of the presence or absence of pheromone, the duration interval of pheromone, the extent of the food supply at the food sources, the size of the colony, and the search strategy employed during foraging, for a given geometry and a finite number of food sources, on the performance of the ant colony. The second and most important objective of this paper is to scientifically study the nature of creativity by modeling synthetic creativity in an ant colony, simulating it on a computer system, and measuring its impact on performance through innovative metric design. The study is motivated by leading thinkers, throughout time, who have strongly emphasized the importance of creativity over intelligence. While an exact and universal definition of creativity is elusive, the important characteristics include originality, intuition, and imagination. Creativity is best understood through its manifestations as novel and radically different ideas that transcend current knowledge and reasoning. In this study, two synthetic creativity traits are introduced into select individual ants of a colony by imparting to them a foraging behavior that is radically different from the normal behavior. Under the first trait, the creative ants choose to ignore the existing pheromone trails and search for food sources. Under the second trait, a creative ant shares its knowledge of food source location, after discovery, with all other creative ants, so that together they can focus solely on discovering new food sources. Analysis of the simulation results reveal that a creative trait coupled with the underlying parameters of the ant colony may cause the foraging completion time metric, i.e. the time to collect food from all sources, to be marginally better or weaker than the normal colony. Contrary to intuition, the completion time metric worsens when, in a fixed sized colony, the number of creative ants, relative to normal ants, becomes excessive. Furthermore, for a given finite geometry, finite food supply, and a fixed ratio of the number of creative to normal ants, the impact of creativity on the foraging performance is dependent on the underlying parameters, and may either continue to improve, remain unchanged, or decrease, as the size of the colony increases. Finally, while the creative ants expend, on an average, two to three times more energy than the normal ants, the ant colony with the creative ants discovers more food sources, in less time and, under certain circumstances, a creative colony discovers all food sources which a normal colony can never achieve.     

社交网络技术在科研教学环境中的应用研究

随着互联网产业的发展、用户规模的扩张和用户行为的变化,以Web2.0为特征的相关需求应运而生。SNS,Social Networking Services,专指旨在帮助人们建立社会性网络的互联网应用服务。依赖其开放性,即时性,互动性,不受时间地点约束等特点获得青睐,很多社交网站如twitter,Facebook,人人网,微博在流量、用户数和市值都有引人瞩目的成绩。社交网络的特性适用于知识的共享,将SNS和E-learning结合产生的面向知识管理的社交网络KMOS系统既使学生在社交网络服务中获取到知识,又打破传统教学受时间、地点的束缚,开放共享知识,一对多的高效传播模式。设计和实现的面向知识管理的社交网络服务拥有的上述优点。使用列表和群组控制多媒体资源的传播范围。分层和插件式设计使系统灵活应对需求改变。使用社交网络技术提高教学质量。     

多特征融合的lncRNA识别与其功能预测

针对传统的基于单一特征的植物lncRNA识别的局限性,提出了融合RNA序列的开放阅读框、二级结构以及k-mers等多特征方法,训练高斯朴素贝叶斯、支持向量机和梯度提升决策树3种经典的分类模型,并实现分类结果的集成,利用交叉验证对模型的性能进行了评估,整体性能优于目前较流行的CPAT、CNCI和PLEK预测软件,在拟南芥数据集上总体的准确率达到了89%。另外,基于内源性竞争规则以及RNA结构信息,分别对lncRNA-microRNA和microRNA-mRNA进行靶向预测、筛选,再通过整合预测数据建立互作网络,并对网络模块中的lncRNA进行功能预测。通过GO术语分析,对与mRNA相关的lncRNA可能参与的生物调控过程进行预测,推测它们的相应功能。     

BERT-TECNN模型的文本分类方法研究

由于Bert-base,Chinese预训练模型参数巨大,在做分类任务微调时内部参数变化较小,易产生过拟合现象,泛化能力弱,且该模型是以字为单位进行的预训练,包含词信息量较少。针对这些问题,提出了BERT-TECNN模型,模型使用Bert-base,Chinese模型作为动态字向量模型,输出包含深度特征信息的字向量,Transformer encoder层再次对数据进行多头自注意力计算,提取特征信息,以提高模型的泛化能力,CNN层利用不同大小卷积核,捕捉每条数据中不同长度词的信息,最后应用softmax进行分类。该模型与Word2Vec+CNN、Word2Vec+BiLSTM、Elmo+CNN、BERT+CNN、BERT+BiLSTM、BERT+Transformer等深度学习文本分类模型在三种数据集上进行对比实验,得到的准确率、精确率、召回率、F1测度值均为最高。实验表明该模型有效地提取了文本中字词的特征信息,优化了过拟合问题,提高了泛化能力。     

人工智能系统故障分析原理研究

为研究未来系统在人工智能控制下的系统故障预测、预防、控制和恢复能力,提出一种基于信息生态方法论的人工智能系统故障分析方法。将研究对象划分为人、功能、自然和智能系统;以智能系统为核心,研究故障信息、知识和智能安全生成原理;论述了基础故障意识、情感和理智的特点。研究表明,系统故障的人工智能分析必须采用信息生态方法论结合安全科学理论进行。分析原理是基于信息生态方法论,考虑基础故障意识、情感与理智,及即时故障语义信息进行的综合决策与反应,以确保系统在规定条件下完成预定功能。     

Real-time pinch detection algorithm: Robust to impulsive noise

A robust pinch detection algorithm which can be implemented in a cheap microprocessor is proposed for the development of a safety feature in the automotive power window system. To solve the problems caused by the performance degradation of a Hall sensor or real driving situations, the proposed algorithm makes use of the H _∞ state estimation technique. The motivation of this approach comes from the advantage that the H _∞ filter can minimize or bound the worst-case estimation error energy for all bounded energy disturbances. Herein, the pinch torque rate estimator is derived from applying the steady-state H _∞ filter to the augmented model, which includes the motor dynamics and an additional torque rate state. Then, to redesign an appropriate estimator for real-time implementation, the torque rate estimate can be calculated more efficiently than the previous method [1]. Experimental results verify that, with a small amount of computation, the proposed pinch detection algorithm provides fast pinch detection performance superior to the existing method. Furthermore, it guarantees robustness against the worst-case measurement noises. Recommended by Editorial Board member Young Soo Suh under the direction of Editor Young Il Lee. Jung-Hoon Park received the B.E. degree in Electronic Engineering in 1996, and the M.S. degree in Electrical and Electronic Engineering from Yonsei University, Seoul, Korea, in 2002. He worked with Samsung Electronics as an Engineer from 1996 to 1999. He is currently pursuing his doctoral degree at Yonsei University. His research interests include robust control and filtering theory, robot vision, and its applications. Won-Sang Ra received the B.E., M.S., and Ph.D. degrees in Electrical and Electronics Engineering from Yonsei University, Seoul, Korea, in 1998, 2000, and 2009, respectively. From March 2000 to February 2009, he was with the Guidance and Control Department of Agency for Defense Development, Daejeon, as a Senior Researcher. Since March 2009, he has been with the School of Mechanical and Control Engineering, Handong Global University, where he is currently a Full-Time Instructor. His main research topic includes the robust filtering theory and its applications to autonomous vehicle guidance and control. Tae-Sung Yoon received the B.E., M.S., and Ph.D. degrees, in Electrical Engineering from Yonsei University, Seoul, Korea, in 1978, 1980, and 1988, respectively. He worked with the Department of Electrical Engineering at the 2nd Naval Academy, Jinhae, Korea, as a member of the teaching staff from 1980 to 1983. He worked with the Department of Electrical Engineering at Vanderbilt University, Nashville, as a Visiting Assistant Professor from 1994 to 1995. Since 1989, he has been with the Department of Electrical Engineering, Changwon National University, Changwon, Korea where he is currently a Professor. His research interests include robust filtering, mobile robotics, and time-frequency signal processing in instrumentation. Jin-Bae Park received the B.E. degree in Electrical Engineering from Yonsei University, Seoul, Korea, in 1977, and the M.S. and Ph.D. degrees in Electrical Engineering from Kansas State University, Manhattan, in 1985, and 1990, respectively. Since 1992, he has been with the Department of Electrical and Electronic Engineering, Yonsei University, where he is currently a Professor. His research interests include robust control and filtering, nonlinear control, mobile robotics, fuzzy logic control, neural networks, genetic algorithms, and Hadamard-transform spectroscopy. He has served as the Director for the Transactions of the Korean Institute of Electrical Engineers (1998–2003) and the Institute of Control, Automation, and Systems Engineers (1999–2003). He is currently the Editor-in-Chief for the International Journal of Control, Automation, and Systems.     

社会Agent的BDO模型

目前MAS中思维状态的研究趋势是在个体模型中加入社会思维属性,研究社会承诺、依赖、联合意图的推理关系。在BDI模型中,以意图为中心的观点不描述社会Agent。该文分析了以竭力为中心的Agent思维状态研究存在的问题,提出MAS的分层模型,并提出以信念、愿望和义务作为基本思维属性（简称BDO）来描述Agent的思维状态和社会属性,给出了BDO逻辑和语义模型,考虑了奖励、惩罚、承诺和承诺解除问题,并对队、组织、组织意图等现象给出了描述。该文力图更自然地描述社会性Agent的思维状态和群体概念,是对Rao和Georgeff提出的BDI模型的改进。最后通过一个例子说明了BDO的表达能力。进一步的工作包括建立更为完善的语义模型、结合各个思维属性的动态修正语义给出BDO Agent的动态模型以及给出基于BDO逻辑框架的Agent／MAS实现结构。     

Testing the usability of smartphone surface gestures on different sizes of smartphones by different age groups of users

This study focuses on testing the usability of 3 different sizes of smartphones by children, adults, and the elderly. Forty-seven participants from each age group were recruited. The research consisted of two stages of experiments. In the first experiment, three existing gestures used by the Android system were evaluated: drag, pinch, and double-touch drag. In the second experiment, two additional re-design gestures were tested: multi-long press and slide down. The users' performances were measured based on their overall time, success time, accuracy, Fitts' Law values, and a subjective questionnaire. Among the five gestures, drag and multi-long press were categorized as the drag gesture, and pinch, double-touch drag and slide down were categorized as the enlarging gesture. The results show that children and adults outperformed the elderly in gesture operation. With respect to smartphone size, the subjects had faster response times and better Fitts' Law values on 6-inch phones. When comparing the subjective preference and the objective use measurement values of the two drag gestures and three enlarging gestures among the different age groups, it was found that children and adults were consistent, whereas the elderly were inconsistent. Overall, in operating the original drag and pinch gestures, the elderly exhibited better objective use measurements, but they preferred the improved gestures, such as slide down.     

Toward a Meta-Model for Computational Engineering

Geometric modeling and finite element analysis have matured in recent decades. Both methods are used extensively in engineering design. However, the link between geometric modeling, physical modeling and finite element analysis is currently cumbersome, error-prone, and ad-hoc. Topological domain modeling provides the missing link. In this paper, we propose a combined topological modeling and finite element modeling method that allows not only topological modeling, but also promotes geometric and physical modeling, by providing a topological base space for the definition of finite element meshes, fields, and the definition and solution of boundary value problems. We call the method the Constructive Topological Domain Method (CTDM). In this method, Primitive Topological Domains (PTDs), each possessing a natural coordinate space, are combined in multiple n-dimensional Cartesian coordinate spaces, called charts, using generalizations of Boolean set operations, to create Constructed Topological Domains (CTDs) capable of acting as the base spaces of fiber bundles. The charts are glued together to create an atlas, within which the CTD is defined. The fiber of the bundle may describe, in addition to geometry, physical fields like density, stress, and temperature. Finite element meshes may be defined upon each of the PTDs from which the CTD is constructed, enabling the definition and solution of boundary value problems, thus avoiding the difficult and messy problem of creating a single finite element mesh to represent the entire CTD. A modified finite element method, to handle the individually meshed PTDs, is described. The boundary conditions may be specified as analytical or as finite element-based fields upon each of the PTDs. The CTDM appears to be a promising approach to robust mathematical and computational modeling of physical objects. Simple examples are presented. ID="A1"Correspondance and offprint requests to: W. Gerstle, Department of Civil Engineering, University of New Mexico, Albuquerque, NM 87131, USA. E-mail: gerstle@unm.edu