自然场景下人脸表情数据集的构建

目前人脸识别研究中表情数据集图像数量较少、表情信息单一，不利于人脸表情识别的研究。本文创建了自然场景下带标签的人脸表情数据集(Facial expression dataset in the wild, FELW)，并对其进行测试。FELW表情数据集包含了多张从互联网上收集的不同的年龄、种族、性别的人脸表情图像，采用适合的方法标注每张图像带有人脸部件的状态标签和表情标签，并引入Kappa一致性检验，提高人脸表情识别率。使用传统方法和深度学习的表情识别方法对数据集进行实验分析，与其他公开的人脸表情数据集相比，FELW数据集具有更多图像和更丰富的表情类别，并包含了两种图像标签有利于表情识别的研究。     

真实感3D人脸表情合成技术研究进展

文中综述了真实感3D人脸表情合成技术的研究进展.根据所采用的技术,将人脸表情合成方法划分为混合样本人脸表情合成、直接表情迁移、基于sketch的人脸表情合成与编辑、基于机器学习的真实感人脸表情合成、高分辨率人脸表情和细节的提取与合成5类,介绍了各类方法的研究现状,并比较了它们的优缺点.最后进行了总结,并对人脸表情合成技术未来的研究方向予以展望.     

基于光流特征与高斯LDA的面部表情识别算法

文中提出了一种人脸表情识别的新方法,该方法采用动态的光流特征来描述人脸表情的变化差异,提高人脸表情的识别率。首先,计算人脸表情图像与中性表情图像之间的光流特征;然后,对传统的线性判断分析方法(Linear Discriminant Analysis,LDA)进行扩展,采用高斯LDA方法对光流特征进行映射,从而得到人脸表情图像的特征向量;最后,设计多类支持向量机分类器,实现人脸表情的分类与识别。在JAFFE和CK人脸表情数据库上的表情识别实验结果表明,该方法的平均识别率比3种对比方法的高出2%以上。     

基于局部SVM分类器的表情识别方法

提出了一种新的视频人脸表情识别方法. 该方法将识别过程分成人脸表情特征提取和分类2个部分,首先采用基于点跟踪的活动形状模型(ASM)从视频人脸中提取人脸表情几何特征;然后,采用一种新的局部支撑向量机分类器对表情进行分类. 在Cohn2Kanade数据库上对KNN、SVM、KNN2SVM和LSVM 4种分类器的比较实验结果验证了所提出方法的有效性.     

人脸样貌独立判别的协作表情识别算法

为了降低样貌、姿态、眼镜以及表情定义不统一等因素对人脸表情识别的影响,提出一种人脸样貌独立判别的协作表情识别算法。首先,采用自动的人脸检测算法定位、对齐视频每帧的人脸区域,并从人脸视频序列中选择峰值表情的人脸;然后,采用峰值人脸与某个表情类内的所有人脸产生表情类内差异人脸信息,并通过计算峰值表情人脸与表情类内差异人脸的差异信息获得协作的表情表示;最终,采用基于稀疏的分类器与表情表示决定每个人脸表情的标签。采用欧美与亚洲人脸的数据库进行仿真实验,结果表明本算法获得了较好的表情识别准确率,对不同样貌、佩戴眼镜的人脸样本也具有较好的识别效果。     

人机交互中的人脸表情识别研究进展

随着人机交互与情感计算技术的快速发展,人脸表情识别已成为人们研究的热点。为了阐明人机交互中人脸表情识别的研究方向及进展,该文从人脸表情数据库、表情特征提取、表情分类方法、鲁棒的表情识别、精细的表情识别、混合表情识别、非基本表情识别等方面对人脸表情识别的研究现状进行了分析。最后总结了人脸表情识别研究的热点及趋势,同时指出了人脸表情识别研究存在的局限性,并对人脸表情识别的发展进行了展望。     

基于拉普拉斯算子的人脸表情克隆优化方法

针对人脸克隆表情的真实性和表情重建的效率,提出了一种基于运动捕捉数据的人脸表情克隆新方法。使用运动捕捉设备捕捉人脸六种基本表情数据,并对数据进行归一化处理,使其位于同一区间范围,保证克隆表情数据的有效性;在表现人脸表情的关键区域选取41个特征点,通过采用凸权值的拉普拉斯变形算法将人脸表情转移到目标模型上,很好保留了人脸表情的细节,克服了均一权值保真度低和余切权值计算不稳定的问题。实验结果表明,该方法合成的克隆表情真实、自然,人脸表情克隆的效率得到了显著提高。     

面部表情的深度卷积级联森林识别

面部表情识别广泛应用于各种研究领域,针对面部表情识别使用深度神经网络方法结构复杂、可解释性差和传统机器学习方法特征提取缺乏多样性、识别率低的问题.提出了一种新的深度卷积级联森林(Deep Convolution Cascade Forest,DCCF)方法用于人脸面部表情识别,该方法通过卷积神经网络深度学习人脸面部显著特征,并采用基于随机森林的级联结构森林逐层学习识别出不同的面部表情特征,提高了人脸表情的识别准确率.DCCF在JAFFE、CK+和Fer2013 3个公开面部表情数据集进行了实验,并对面部表情提取的5种特征和7种分类方法进行了比较分析,结果显示DCCF在对比的算法中人脸表情识别性能最好,3个数据集的准确率分别达到91.4%,98.7%,71.6%.     

人脸性别约束下的深度随机森林表情识别

由于人脸表情类内变化和类间干扰因素的存在,人脸表情识别仍面临着巨大挑战.提出一种基于性别条件约束随机森林的深度人脸表情识别方法,解决人脸表情识别中噪声、性别等变化和干扰问题.首先,采用深度多示例学习方法提取鲁棒性人脸特征,解决人脸光照、遮挡和低分辨率等图像变化问题;其次,采用性别条件随机森林分类方法进行人脸表情分类器设计,解决人脸性别因素干扰问题.在公开的CK+、BU-3DEF、LFW人脸表情数据库上进行广泛实验结果表明:所提出方法在3大人脸数据库上分别达到了98.83%、90%、60.58%的识别率,与先进方法相比具有更好的性能和鲁棒性.另外,与其他先进的深度学习方法(需要大量训练数据库)相比,所提出方法只需要小量训练样本就能达到较好效果.     

基于小尺度核卷积的人脸表情识别

针对现有表情识别方法中网络泛化能力差以及网络参数多导致计算量大的问题,提出一种利用小尺度核卷积的人脸表情识别方法。采用多层小尺度核卷积块代替大卷积核减少参数量,结合最大池化层提取面部表情图像特征,利用Softmax分类器对不同表情进行分类,并在相同感受野下增加网络深度避免特征丢失。实验结果表明,与FER2013 record、DNNRL等方法相比,该方法的人脸表情识别率更高,能有效实现人脸表情的准确分类。     

Probability representations of fuzzy systems

1 Background of birth of fuzzy systems As is well-known, just considering the great uncertainties of many systems, Zadeh put forward the notion of fuzzy sets and advanced the idea of the fuzzy reasoning by means of fuzzy sets which could describe a system approximately[1]. Those systems that are constructed on the basis of fuzzy reasoning are called fuzzy systems in general[2―4]. The research on fuzzy systems has attracted broad attention[5―7]. For instance, universal ap- proximation proper…     

结合K-D树和Shell的快速动态等值面光线跟踪法

目的 K-D树和Shell常被用于加速等值面光线跟踪法。如果Isovalue保持不变时,Shell方法效率更高,否则Shell方法需要重构Shell,而K-D树方法速度更快。提出一种结合K-D树和Shell的动态等值面光线跟踪方法,其关键是如何实现两者的平稳切换。方法首先改进基于K-D树的等值面光线跟踪算法,使得该方法在绘制过程中渐进地构建Shell数据结构。在Isovalue发生变化后,首先使用改进的基于K-D树的等值面光线跟踪算法进行绘制,并在绘制过程中渐进地构建新的Shell数据结构,从而平滑地过渡到绘制效率更高的基于Shell的等值面光线跟踪算法。结果实验中,在Isovalue动态变化时本文方法的效率接近K-D树方法;但用户在进行缩放、旋转等操作时,本文方法能达到Shell方法的速度。实验结果表明本文方法结合了两者的优点。结论提出了一种快速Isovalue光线跟踪算法,综合运用K-D树和Shell两种数据结构,在Isovalue保持不变和动态变化两者场合都实现了较高的绘制速度。     

A distance based clustering method for arbitrary shaped clusters in large datasets

Clustering has been widely used in different fields of science, technology, social science, etc. Naturally, clusters are in arbitrary (non-convex) shapes in a dataset. One important class of clustering is distance based method. However, distance based clustering methods usually find clusters of convex shapes. Classical single-link is a distance based clustering method, which can find arbitrary shaped clusters. It scans dataset multiple times and has time requirement of O(n²), where n is the size of the dataset. This is potentially a severe problem for a large dataset. In this paper, we propose a distance based clustering method, l-SL to find arbitrary shaped clusters in a large dataset. In this method, first leaders clustering method is applied to a dataset to derive a set of leaders; subsequently single-link method (with distance stopping criteria) is applied to the leaders set to obtain final clustering. The l-SL method produces a flat clustering. It is considerably faster than the single-link method applied to dataset directly. Clustering result of the l-SL may deviate nominally from final clustering of the single-link method (distance stopping criteria) applied to dataset directly. To compensate deviation of the l-SL, an improvement method is also proposed. Experiments are conducted with standard real world and synthetic datasets. Experimental results show the effectiveness of the proposed clustering methods for large datasets.     

A dissipative exponentially-fitted method for the numerical solution of the Schrödinger equation and related problems

In this paper a dissipative exponentially-fitted method for the numerical integration of the Schrödinger equation and related problems is developed. The method is called dissipative since is a nonsymmetric multistep method. An application to the the resonance problem of the radial Schrödinger equation and to other well known related problems indicates that the new method is more efficient than the corresponding classical dissipative method and other well known methods. Based on the new method and the method of Raptis and Cash a new variable-step method is obtained. The application of the new variable-step method to the coupled differential equations arising from the Schrödinger equation indicates the power of the new approach.     

分布式视频编码中相关噪声分布的非参数估计

目的 针对分布式视频编码系统中相关噪声（CN）分布难以准确模拟的问题,提出了一种CN的非参数估计方法。方法 根据CN分布的特点,提出CN的非参数估计方法,建立了基于最优窗宽的核密度估计-均匀分布模型（KDEUDM）,比较了变换域Wyner-Ziv（TDWZ）系统中CN的参数估计法和非参数估计法所建立的噪声模型对系统性能的影响。结果 实验结果表明,非参数估计方法能较准确地模拟CN的分布,与参数估计法相比,用非参数估计法建立的噪声模型能使WZ帧编码在高码率下最高能节约10%的码率。结论 非参数估计法是TDWZ系统中有效的相关噪声估计方法。     

Model-based prediction error uncertainty estimation for k-nn method

The k-nearest neighbour estimation method is one of the main tools used in multi-source forest inventories. It is a powerful non-parametric method for which estimates are easy to compute and relatively accurate. One downside of this method is that it lacks an uncertainty measure for predicted values and for areas of an arbitrary size. We present a method to estimate the prediction uncertainty based on the variogram model which derives the necessary formula for the k-nn method. A data application is illustrated for multi-source forest inventory data, and the results are compared at pixel level to the conventional RMSE method. We find that the variogram model-based method which is analytic, is competitive with the RMSE method.     

A Novelty-based Clustering Method for On-line Documents

In this paper, we describe a document clustering method called novelty-based document clustering. This method clusters documents based on similarity and novelty. The method assigns higher weights to recent documents than old ones and generates clusters with the focus on recent topics. The similarity function is derived probabilistically, extending the conventional cosine measure of the vector space model by incorporating a document forgetting model to produce novelty-based clusters. The clustering procedure is a variation of the K-means method. An additional feature of our clustering method is an incremental update facility, which is applied when new documents are incorporated into a document repository. Performance of the clustering method is examined through experiments. Experimental results show the efficiency and effectiveness of our method.     

三对角线性方程组的分布式并行算法

文中回顾了Ｍｉｃｈｉｅｌｓｅ＆Ｖｏｒｓｔ算法，分析了影响Ｍ＆Ｖ．算法并行效率的主要因素，根据分布治之思想提出了一种求解三对角方程组的并行算法，新算法通信建立次数的Ｍ＆Ｖ算法的５０％，数据传输量为其３３％，最后的工作站网络环境下实现了新算法，就并行效率与Ｍ＆Ｖ算法进行了比较，结果表明在由６台工作站的组成的网络中新算法必能提高可达到４０％。     

Line Segment Intersection Testing

A method for accurately determining whether two given line segments intersect is presented. This method uses the standard floating-point arithmetic that conforms to IEEE 754 standard. If three or four ending points of the two given line segments are on a same vertical or horizontal line, the intersection testing result is obtained directly. Otherwise, the ending points and their connections are mapped onto a 3×3 grid, and the intersection testing falls into one of the five testing classes. The intersection testing method is based on our method for floating-point dot product summation, whose error bound is 1ulp. Our method does not have the limitation in the method of Gavrilova and Rokne (2000) that the product of two floating-point numbers is calculated by a twice higher precision floating-point arithmetic than that of the multipliers. Furthermore, this method requires less than one-fifth of the running time used by the method of Gavrilova and Rokne (2000), and our new method for calculating the sign of a sum of n floating-point numbers requires less than one-fifteenth of the running time used by ESSA.     

Parallel solution of dense linear systems using diagonalization methods

We study the parallel implementation of two diagonalization methods for solving dense linear systems: the well known Gauss-Jordan method and a new one introduced by Huard. The number of arithmetic operations performed by the Huard method is the same as for Gaussian elimination, namely 2n ³/3, less than for the Jordan method, namely n ³. We introduce parallel versions of these methods, compare their performances and study their complexity. We assume a shared memory computer with a number of processors p of the order of n, the size of the problem to be solved, We show that the best parallel version for Jordan's method is by rows whereas the best one for Huard's method is by columns. Our main result states that for a small number of processors the parallel Huard method is faster than the parallel Jordan method and slower otherwise. The separation is obtained for p = 0.44n.