二值化的SIFT特征描述子及图像拼接优化

目的 针对SIFT算法计算复杂度高、存储开销大和近几年提出的BRIEF（binary robust independent elementary features）、ORB（oriented BRIEF）、BRISK（binary robust invariant scalable keypoints）和FREAK（fast retina keypoint）等二进制描述子可区分性弱和鲁棒性差的问题,提出基于SIFT的二进制图像局部特征描述子。方法 首先,对传统SIFT的特征空间和特征向量分布在理论和实验上进行分析,在此基础上结合二进制特征描述子的优势对SIFT进行改进。不同于传统的二进制特征描述子,本文算法对传统SIFT特征向量在每一维上的分量进行排序后,以该特征向量的中值作为量化阈值,将高维浮点型SIFT特征向量转化成位向量得到二进制特征描述子。并使用易于计算的汉明距离代替欧氏距离度量特征点间的相似性以提高匹配效率。然后,在匹配阶段将二进制特征描述子分为两部分并分别对其进行匹配,目的是通过初匹配剔除无效匹配特征点来进一步缩短匹配时间。最后,对提出的量化算法的可区分性及鲁棒性进行验证。结果 该量化算法在保持SIFT的较强的鲁棒性和可区分性的同时,达到了低存储、高匹配效率的要求,解决了SIFT算法的计算复杂度高、二进制描述子鲁棒性和可区分性差的问题。此外,在匹配阶段平均剔除了77.5%的无效匹配特征点,减少了RANSAC（random sample consensus）的迭代次数。结论 本文提出的量化算法可用于快速匹配和快速图像拼接中,提高匹配和拼接效率。     

Hand posture and gesture recognition technology

Hand gestures that are performed by one or two hands can be categorized according to their applications into different categories including conversational, controlling, manipulative and communicative gestures. Generally, hand gesture recognition aims to identify specific human gestures and use them to convey information. The process of hand gesture recognition composes mainly of four stages: hand gesture images collection, gesture image preprocessing using some techniques including edge detection, filtering and normalization, capture the main characteristics of the gesture images and the evaluation (or classification) stage where the image is classified to its corresponding gesture class. There are many methods that have been used in the classification stage of hand gesture recognition such as Artificial Neural Networks, template matching, Hidden Markov Models and Dynamic Time Warping. This exploratory survey aims to provide a progress report on hand posture and gesture recognition technology.     

一种抗视角变换的SURF匹配算法

针对传统的SURF匹配算法在图像搜索、物体识别中对视角变化大的图像特征点无法匹配的缺陷,提出了一种改进的SURF图像匹配算法。该算法利用模拟图序列的概念,即由原始图像模拟各种可能的视角变换生成的一组图像序列,对图像序列中的每幅图像进行SURF特征点提取;剔除冗余特征点后将模拟图序列中的特征点映射到原始图像中,这一过程增加了原始图像的特征点数量,特征点描述向量改用原始图像和多个模拟图的SURF特征描述向量共同表示;最后采用欧氏距离准则,最终得到两幅图像的正确匹配对。实验结果表明,本算法增加了特征点匹配对数,改善了视角变换过大(大于60°)使得图像无法匹配的情况。     

The IT Value Hierarchy: Using Maslow's Hierarchy of Needs as a Metaphor for Gauging the Maturity Level of Information Technology Use within Competitive Organizations

Abstract

The purpose of this paper is to present a hierarchy of progressive IT maturity using Maslow's Hierarchy of Needs as a metaphor for articulating the increasing value that can be derived by the application of Information Technology within competitive organizations. The authors refer to this maturity model as the IT Value Hierarchy. Each level of the IT Value Hierarchy is described using examples and comparisons to Maslow's Hierarchy. The model can be used by IT executives as a framework for better explaining and discussing the value of increasingly sophisticated Information Technology use within the enterprise.     

一种改进的SIFT特征匹配算法

针对尺度不变特征变换(SIFT)特征匹配算法存在计算量大、复杂度高的问题,提出一种基于图像Radon变换的改进SIFT特征匹配算法。改进算法在图像的SIFT特征点采样区域内作d条不同方向的直线,以d条直线上的图像Radon变换作为SIFT特征向量描述符,降低SIFT特征向量的维数,从而提高特征匹配效率。实验结果表明,改进算法具有较高的匹配精度和较少的匹配时间,适用于虚拟场景漫游或目标识别等实时性要求较高的系统。     

字典重建和空间分布关系约束下的特征选择与图像拼接

目的 针对大型图像检索领域中,复杂图像中SIFT特征描述子的冗余和高维问题,提出了一种基于字典重建和空间分布关系约束的特征选择的方法,来消除冗余特征并保留最具表现力的、保留原始空间结构性的SIFT特征描述子。方法 首先,实验发现了特征选择和字典学习方法在稀疏表示方面的内在联系,将特征选择问题转化为字典重构任务;其次,在SIFT特征选择问题中,为了保证特征空间中特征的鲁棒性,设计了新型的字典学习模型,并采用模拟退火算法进行迭代求解;最后,在字典学习的过程中,加入熵理论来约束特征的空间分布,使学习到的特征描述子能最大限度保持原始SIFT特征空间的空间拓扑关系。结果 在公开数据集Holiday大型场景图片检索数据库上,通过与国际公认的特征选择方法进行实验对比,本文提出的特征选择方法在节省内存空间和提高时间效率（30%~ 50%）的同时,还能保证所筛选的特征描述子的检索准确率比同类特征提高8%~ 14.1%;在国际通用的大型场景图片拼接数据库IPM上,验证本文方法在图像拼接应用中特征提取和特征匹配上的有效性,实验表明本文方法能节省（50% ~70%）图像拼接时间。结论 与已有的方法比较,本文的特征选择方法既不依赖训练数据集,也不丢失重要的空间结构和纹理信息,在大型图像检索、图像拼接领域和3D检索领域中,能够精简特征,提高特征匹配效率和准确率。     

ICT supported healthy lifestyle interventions: design lessons

Healthy living is increasingly urgent, since it may help curb our escalating health care costs. ICT (Information and Communication Technology) offers healthy lifestyle support opportunities. Given the dose–response relation between health behaviors and -benefits, an important challenge is: how to motivate participants to large health behavior improvements? We conducted two design and test iterations to add ICT support to traditional face to face coach formats: first web based eDashboarding. Next: integrating existing best of breed mobile smart phone apps (mApps) in the coach relationship. We extract several design lessons regarding: how to design for a motivational virtuous circle, and when do Web or mApps add benefits for users? An important finding is that eCoach solutions add more value when part of an overall health coach relationship. For the future, we anticipate more intelligent mobile applications for health behavior tracking and feedback, plus an increasing role for mApps in health provisioning.     

高斯二阶差分特征算子在图像拼接中的应用

为了将同一场景中具有重叠区域序列的图像快速准确合成一幅具有宽视角、高分辨率的图像,提出了基于高斯二阶差分(D2oG)特征检测算子的SIFT算法.采用高斯二阶差分(D2oG)金字塔的过零点检测提取图像尺度不变特征点,并选用RANSAC算法对特征点匹配对进行提纯,在此基础上计算不变换矩阵H,最后,用渐进渐出平滑算法完成图像的无缝拼接.实验中分别采用所提方法和SIFT算法对具有典型变换的4种图像进行拼接与测试,结果表明:所提方法提取的匹配点数、拼接所消耗时间明显低于采用SIFT算法,同时匹配效率也高于后者.此方法降低了运算复杂度的同时,图像拼接实时性也得到提高.     

An effective head pose estimation approach using Lie Algebrized Gaussians based face representation

The accuracy of head pose estimation is significant for many computer vision applications such as face recognition, driver attention detection and human-computer interaction. Most appearance-based head pose estimation works typically extract the low-dimensional face appearance features in some statistic subspaces, where the subspaces represent the underlying geometry structure of the pose space. However, there is an open problem, namely, how to effectively represent appearance-based subspace face for the head pose estimation problem. To address the problem, this paper proposes a head pose estimation approach based on the Lie Algebrized Gaussians (LAG) feature to model the pose characteristic. LAG is built on Gaussian Mixture Models (GMM), which actually not only models the distribution of local appearance features, but also captures the Lie group manifold structure of the feature space. Moreover, to keep multi-resolution structure information, LAG is operated on many subregions of the image. As a result, these properties of LAG enable it to effectively model the structure of subspace face which can lead to powerful discriminative ability for head pose estimation. After representing subspace face using the LAG, we treat the head pose estimation as a classification problem. The within-class covariance normalization (WCCN) based Support Vector Machine (SVM) classifier is employed to achieve robust performance as WCCN could reduce the within-class variabilities of the same pose. Extensive experimental analysis and comparison with both traditional and state-of-the-art algorithms on two challenging benchmarks demonstrate the effectiveness of our approach.     

多时相遥感影像的深度卷积匹配算法研究

针对传统人工设计的特征描述符在面对几何差异和辐射差异较大、地物变化明显的多时相遥感影像时,容易造成大量误匹配问题,提出一种适用于遥感影像的双筛选双约束的卷积神经网络匹配算法,通过CNN提取影像深层语义特征图,在特征图中筛选同时满足优先最大原则和精确极值原则的关键点,并在对应位置提取512维描述符。在特征匹配阶段,通过由粗到精的反向匹配约束和RANSAC约束相结合的策略,保证了误匹配点有效剔除的同时保留一定数量的正确匹配点对,提高了匹配的精度。实验结果表明：该方法的匹配数量适中,正确匹配率较实验选取的其他算法相比有所提高,影像配准误差减小2个像素左右,匹配运行时间提升明显,匹配效果良好并且具有一定的适应性和鲁棒性。     

Fusion visualization for fluid dynamics in blood vessel

The speed up of supercomputers has increased the complexity of simulations. To analyze such kind of data, new types of visualization software are needed. As one of approach for meeting this requirement, we are developing the “Fusion Visualization” in a project sponsored by the Japan Science and Technology Agency (JST). It can execute fused visualization of simulation data combining both volume and surface rendering. The overall concept was reported last year at the AROB 18th International Symposium in Korea. In this work, we are reporting the ongoing research with an application example related to blood flow simulation.     

基于距离融合的图像特征点匹配方法

针对ORB算法中特征点缺乏尺度不变性导致算法误匹配率高,以及二进制鲁棒独立基本特征（BRIEF）算法的描述子易受噪声影响的问题,提出了改进的特征点匹配方法。采用加速的具有鲁棒性的特征（SURF）算法进行特征点提取,利用带有方向信息的BRIEF算法进行特征点描述;在特征点邻域内选取随机点对,并对随机点对的灰度大小比较和相似度比较分别进行编码,采用汉明距离计算两种编码的差异;利用自适应加权融合的方式实现特征点相似性距离度量。实验结果表明,改进方法对于尺度变化、光照变化以及模糊变化的图像具有更好的适应性,与传统ORB特征点匹配方法相比能够获得更高的特征点正确匹配率,且该特征点匹配方法可用于改善图像拼接的性能。     

具有抗仿射性的局部特征点匹配方法

针对现有局部特征匹配算法对具有仿射性的图像匹配效果欠佳、耗时较长,以及随机采样一致性（RANSAC）算法对仿射性图像匹配得不到较好的参数模型等问题,提出一种具有抗仿射性的A-AKAZE（Affine Accelerated KAZE）算法,并用向量场一致性来筛选内点。首先利用非线性函数构建尺度空间,然后借助Hessian矩阵检测特征点,并以特征点为中心选取合适的区域作为特征采样窗口;再把特征采样窗口在经纬度上进行投影以模拟不同角度对图像的影响,随后在投影区域中提取具有抗仿射性的A-MLDB（Affine Modified-Local Difference Binary）描述符;最后利用向量场一致性算法提取内点。实验结果表明：A-AKAZE算法的正确匹配率相较于AKAZE算法提高了20%以上,与AKAZE+RANSAC算法相较提升了15%左右,与ASIFT（Affine Scale-Invariant Feature Transform）算法相比提高了10%左右,相比ASIFT+RANSAC算法提高了5%;而且该算法的匹配速度远高于AKAZE+RANSAC、ASIFT和ASIFT+RANSAC算法。     

基于改进ORB的视觉里程计算法

ORB-SLAM2因高斯金字塔离散、量化尺度的因素,易造成尺度量化误差.针对这一缺陷,提出孪生过滤器算法.通过在对数极坐标系下进行描述子构造,减少其量化误差,并在同层金字塔中构造笛卡尔坐标系下孪生描述子,利用该描述子距离实现过滤,从而提高特征点的尺度不变性,增强其匹配准确率.同时,针对ORB-SLAM2的四叉树算法过度追求离散度而忽略特征点质量问题这一情况,提出深度有限四叉树算法.利用特征点提取阈值以及特征点所在金字塔层进行自适应深度阈值设置,减小弱特征点区域的划分次数,从而减少弱特征点提取数目.实验表明,所提出算法能够有效提高特征点离散度、正确匹配特征点数目和匹配精度,具有更高的轨迹精度.     

一种改进的OSID的图像匹配算法

针对OSID在构建描述符时未考虑一个特征点的图像块里存在其他特征点,以及生成直方图描述子匹配速度较慢的问题,提出一种基于OSID的改进二进制描述符。在OSID描述符构建的过程中,扇形个数m的选择是固定的,因此提出当一个特征点的图像块里有多个特征点时,尝试将m的值自适应,丰富描述子所包含的信息,提高算法的正确匹配率;并将OSID最后生成的直方图描述子编码成二进制描述子,使用汉明距离代替欧氏距离进行图像匹配,提高算法的匹配速度。在标准数据集上进行测试,结果表明在复杂的视点变化、图像模糊和JPEG压缩等场景下,改进OSID的匹配精度优于同类描述符以及原算法。     

High-MCC Functions in the Linux Kernel

McCabe’s Cyclomatic Complexity (MCC) is a widely used metric for the complexity of control flow. Common usage decrees that functions should not have an MCC above 50, and preferably much less. However, the Linux kernel includes more than 800 functions with MCC values above 50, and over the years 369 functions have had an MCC of 100 or more. Moreover, some of these functions undergo extensive evolution, indicating that developers are successful in coping with the supposed high complexity. Functions with similarly high MCC values also occur in other operating systems and domains, including Windows. For example, the highest MCC value in FreeBSD is 1316, double the highest MCC in Linux. We attempt to explain all this by analyzing the structure of high-MCC functions in Linux and showing that in many cases they are in fact well-structured (albeit we observe some cases where developers indeed refactor the code in order to reduce complexity). Moreover, human opinions do not correlate with the MCC values of these functions. A survey of perceived complexity shows that there are cases where high MCC functions were ranked as having a low complexity. We characterize these cases and identify specific code attributes such as the diversity of constructs (not only a switch but also ifs) and nesting that correlate with discrete increases in perceived complexity. These observations indicate that a high MCC is not necessarily an impediment to code comprehension, and support the notion that complexity cannot be fully captured using simple syntactic code metrics. In particular, we show that regularity in the code (meaning repetitions of the same pattern of control structures) correlates with low perceived complexity.     

一种结合SIFT和对应尺度LTP综合特征的图像匹配算法

由于SIFT特征是一种性能良好的局部特征,常被广泛应用于图像匹配,但SIFT特征点有128维描述符,所以具有匹配复杂度高和计算量大等缺点。为了提高图像匹配效率,研究了一种新的图像匹配方法。该方法通过构建尺度空间、检测极值点、确定关键点等步骤生成SIFT关键特征点;然后利用特征点周围邻域点的旋转不变LTP特征和相对灰度直方图来描述,替代传统SIFT特征点的128维描述,图像匹配过程中使用街区距离代替欧氏距离;最后利用光照变化、模糊变化、尺度和旋转综合变化三组图像进行算法仿真匹配实验。实验结果表明,本算法在图像尺度、旋转、光照变化条件下具有更高的匹配精确度,并且有效地提高了图像的匹配速度。     

The Impact of Perceived Work Complexity and Shared Leadership on Team Performance of IT Employees of South African Firms

ABSTRACT

This article investigates how work complexity, as perceived by Information Technology (IT) employees, influences team performance, and examines the role played by shared leadership. The findings of this study that collected data from 204 IT employees in South Africa show that perceived work complexity negatively predicts team performance while shared leadership negatively predicts perceived work complexity and positively predicts team performance.     

改进SURF特征的维吾尔文复杂文档图像匹配检索

针对图像局部特征的词袋模型（Bag-of-Word,BOW）检索研究中聚类中心的不确定性和计算复杂性问题,提出一种由不同种类的距离进行相似程度测量的检索和由匹配点数来检索的方法。这种方法首先需要改进文档图像的SURF特征,有效降低特征提取复杂度;其次,对FAST+SURF特征实现FLANN双向匹配与KD-Tree+BBF匹配,在不同变换条件下验证特征鲁棒性;最后,基于这两种检索方法对已收集整理好的各类维吾尔文文档图像数据库进行检索。实验结果表明:基于距离的相似性度量复杂度次于基于匹配数目的检索,而且两种检索策略都能满足快速、精确查找需求。