Using a Visual Attention Model to Improve Gaze Tracking Systems in Interactive 3D Applications

This paper introduces the use of a visual attention model to improve the accuracy of gaze tracking systems. Visual attention models simulate the selective attention part of the human visual system. For instance, in a bottom‐up approach, a saliency map is defined for the image and gives an attention weight to every pixel of the image as a function of its colour, edge or intensity. Our algorithm uses an uncertainty window, defined by the gaze tracker accuracy, and located around the gaze point given by the tracker. Then, using a visual attention model, it searches for the most salient points, or objects, located inside this uncertainty window, and determines a novel, and hopefully, better gaze point. This combination of a gaze tracker together with a visual attention model is considered as the main contribution of the paper. We demonstrate the promising results of our method by presenting two experiments conducted in two different contexts: (1) a free exploration of a visually rich 3D virtual environment without a specific task, and (2) a video game based on gaze tracking involving a selection task. Our approach can be used to improve real‐time gaze tracking systems in many interactive 3D applications such as video games or virtual reality applications. The use of a visual attention model can be adapted to any gaze tracker and the visual attention model can also be adapted to the application in which it is used.     

一种基于聚类的核向量机参数C选择算法

核向量机可以高效学习大样本数据集,却有泛化能力低的缺陷.针对已有参数C选择算法缺乏启发性以及选取困难的不足,本文在分析了核聚类算法和距离比较算法的基础之上,提出基于核聚类的相对距离比较方法,该算法利用核聚类算法在特征空间对样本点进行聚类分簇,然后根据样本点到簇心相对距离的比值,得到参数C.本文在理论和实验两个方面,证明该算法有效地选择参数C,从而提高核支持向量机算法的泛化能力.     

基于多例学习的Web图像聚类

在图像分类和自动标注系统中,多例学习(MIL)是研究的热点.目前MIL中的算法多为监督学习方法.针对非监督学习,在基于EM算法和启发式迭代优化算法的框架下,提出了6种多例聚类算法,并通过它们对来自于真实Web环境下的图像进行聚类以分析用户的搜索兴趣.由于一幅图像含有若干个区域,每个区域可被看为一个样例,属于同一个图像的区域则组成一个包.因此如何理解图像语义内容的问题即转化为多例学习.在多例学习的经典数据集MUSK数据和来自于Web图像集上的比较实验表明,提出的多例聚类算法具有优良的聚类性能.     

基于ISODATA法的冬小麦产量分级监测预报

以江苏省姜堰市为例,进行了基于TM卫星遥感技术和小麦估产模型的冬小麦产量监测研究。在利用GPS实地采样调查和建立解译标志的基础上,通过影像校正、采用优化的ISODATA分类方法,结合人机交互式判读解译等操作,将样点的作物信息数据贯穿到整个校验分类过程中,信息解译精度在90%以上。利用分类提取的冬小麦数据,反演叶面积指数、生物量信息等,结合冬小麦估产模型,计算单点产量信息,经过线性转换,对整个区域的冬小麦产量进行监测预报,并制作了冬小麦产量分级专题图。     

基于维基百科的自动词义消歧方法

针对传统词义消歧仅基于上下文语境而导致准确率低的问题,提出一种多策略的无监督自动词义消歧方法。利用从维基百科在线中提炼出的丰富语义知识,线性融合上下文语境、背景知识和语义信息3大特征,根据逻辑回归算法学习各特征的权重,选取最大融合值所对应的候选项作为最优词义。在SENSEVAL数据集上取得了85．50％的平均准确率,验证了该方法的有效性。     

基于无指导学习的微博评论分析方法

该文以一种有效的方法寻找出有价值的微博评论,这对于读者更高效地阅读评论,为舆情分析、文本挖掘等任务提供支持,均具有重要的应用价值。针对微博及其评论文本短小、内容发散等特点,该文提出一种基于无指导学习的微博评论分析方法,该方法通过互联网搜索引擎扩展微博文本,基于相关性计算自动构造正负训练用例,生成特定的某条微博评论分类模型,通过该模型对评论的价值性进行评估。实验结果表明,该方法能够比较好地识别出评论的价值。     

基于概率的无监督缺陷预测方法

软件缺陷预测能够提高软件开发和测试的效率,保障软件质量。无监督缺陷预测方法具有不需要标签数据的特点,从而能够快速应用于工程实践中。提出了基于概率的无监督缺陷预测方法—PCLA,将度量元值与阈值的差值映射为概率,使用概率评估类存在缺陷的可能性,然后再通过聚类和标记来完成缺陷预测,以解决现有无监督方法直接根据阈值判断时对阈值比较敏感而引起的信息丢失问题。将PCLA方法应用在NetGen和Relink两组数据集,共7个软件项目上,实验结果表明PCLA方法在查全率、查准率、F-measure上相对现有无监督方法分别平均提升4.1%、2.52%、3.14%。     

A fuzzy reasoning and fuzzy-analytical hierarchy process based approach to the process of railway risk information: A railway risk management system

Risk management is becoming increasingly important for railway companies in order to safeguard their passengers and employees while improving safety and reducing maintenance costs. However, in many circumstances, the application of probabilistic risk analysis tools may not give satisfactory results because the risk data are incomplete or there is a high level of uncertainty involved in the risk data. This article presents the development of a risk management system for railway risk analysis using fuzzy reasoning approach and fuzzy analytical hierarchy decision making process. In the system, fuzzy reasoning approach (FRA) is employed to estimate the risk level of each hazardous event in terms of failure frequency, consequence severity and consequence probability. This allows imprecision or approximate information in the risk analysis process. Fuzzy analytical hierarchy process (fuzzy-AHP) technique is then incorporated into the risk model to use its advantage in determining the relative importance of the risk contributions so that the risk assessment can be progressed from hazardous event level to hazard group level and finally to railway system level. This risk assessment system can evaluate both qualitative and quantitative risk data and information associated with a railway system effectively and efficiently, which will provide railway risk analysts, managers and engineers with a method and tool to improve their safety management of railway systems and set safety standards. A case study on risk assessment of shunting at Hammersmith depot is used to illustrate the application of the proposed risk assessment system.     

Efficient Vector Quantization Using the WTA-Rule with Activity Equalization

We propose a new algorithm for vector quantization, the Activity Equalization Vector quantization (AEV). It is based on the winner takes all rule with an additional supervision of the average node activities over a training interval and a subsequent re-positioning of those nodes with low average activities. The re-positioning is aimed to both an exploration of the data space and a better approximation of already discovered data clusters by an equalization of the node activities. We introduce a learning scheme for AEV which requires as previous knowledge about the data only their bounding box. Using an example of Martinetz et al. [1], AEV is compared with the Neural Gas, Frequency Sensitive Competitive Learning (FSCL) and other standard algorithms. It turns out to converge much faster and requires less computational effort.     

基于统计方法的中文姓名识别

本文介绍一个中文姓名的自动识别系统,该系统使用从姓名样本库和真实文本语料库中得到的大量统计数据,以提高系统识别性能。我们从1994年人民日报中随机抽取100篇文章作为测试样本,实验结果表明,准确率和召回率可同时达到90%以上。