提取多尺度光照不变量的人脸识别

光照变化条件下的人脸图像识别一直以来都是图像处理中的热点和难点问题,为了提高人脸图像的识别率,提出了一种用于非均匀光照条件下人脸识别的算法.利用对数及二维小波变换的多尺度特性提取出人脸的光照不变量,然后运用PCA+LDA方法进行人脸特征提取,并采用基于欧氏距离的最近邻分类器进行识别.通过Matlab编程实验,在Yale B人脸库中达到了较高的识别率.     

A test for stability robustness of linear time‐varying systems utilizing the linear time‐invariant ν‐gap metric

A stability robustness test is developed for internally stable, nominal, linear time‐invariant (LTI) feedback systems subject to structured, linear time‐varying uncertainty. There exists (in the literature) a necessary and sufficient structured small gain condition that determines robust stability in such cases. In this paper, the structured small gain theorem is utilized to formulate a (sufficient) stability robustness condition in a scaled LTI ν‐gap metric framework. The scaled LTI ν‐gap metric stability condition is shown to be computable via linear matrix inequality techniques, similar to the structured small gain condition. Apart from a comparison with a generalized robust stability margin as the final part of the stability test, however, the solution algorithm implemented to test the scaled LTI ν‐gap metric stability robustness condition is shown to be independent of knowledge about the controller transfer function (as opposed to the LMI feasibility problem associated with the scaled small gain condition which is dependent on knowledge about the controller). Thus, given a nominal plant and a structured uncertainty set, the stability robustness condition presented in this paper provides a single constraint on a controller (in terms of a large enough generalized robust stability margin) that (sufficiently) guarantees to stabilize all plants in the uncertainty set. Copyright © 2008 John Wiley & Sons, Ltd.     

Measuring robustness of computing systems

System builders are becoming increasingly interested in robust design. We believe that a methodology for generating robustness metrics will help the robust design research efforts and, in general, is an important step in the efforts to create robust computing systems. The purpose of the research in this paper is to quantify the robustness of a resource allocation, with the eventual objective of setting a standard that could easily be instantiated for a particular computing system to generate a robustness metric. The paper exposes how not considering the uncertainties can result in gross overestimation of the system capacity, and shows a method for reducing the impact of uncertainty, and even use it to our advantage if the uncertainties are correlated. We present our theoretical foundation for a robustness metric and give its instantiation for a particular system.     

Controller certification

A modern problem from aerospace control involves the certification of a large set of potential controllers with either a single plant or a fleet of potential plant systems, with both plants and controllers being MIMO and, for the moment, linear. Experiments on a limited number of controller/plant pairs should establish the stability and a certain level of margin of the complete set. We consider this certification problem for a set of controllers and provide algorithms for selecting an efficient subset for testing. This is done for a finite set of candidate controllers and, at least for SISO plants, for compact infinite set. In doing this, the ν-gap metric will be the main tool. Computational examples are given, including one of certification of an aircraft engine controller. The overarching aim is to introduce truly MIMO margin calculations and to understand their efficacy in certifying stability over a set of controllers and in replacing legacy single-loop gain and phase margin calculations.     

Weakly supervised temporal action localization with proxy metric modeling

Temporal localization is crucial for action video recognition. Since the manual annotations are expensive and time-consuming in videos, temporal localization with weak video-level labels is challenging but indispensable. In this paper, we propose a weakly-supervised temporal action localization approach in untrimmed videos. To settle this issue, we train the model based on the proxies of each action class. The proxies are used to measure the distances between action segments and different original action features. We use a proxy-based metric to cluster the same actions together and separate actions from backgrounds. Compared with state-of-the-art methods, our method achieved competitive results on the THUMOS14 and ActivityNet1.2 datasets.     

Computing the Kantorovich Distance for Images

Computing the Kantorovich distance for images is equivalent to solving a very large transportation problem. The cost-function of this transportation problem depends on which distance-function one uses to measure distances between pixels.In this paper we present an algorithm, with a computational complexity of roughly order (N2), where N is equal to the number of pixels in the two images, in case the underlying distance-function is the L1-metric, an approximation of the L2-metric or the square of the L2-metric; a standard algorithm would have a computational complexity of order (N3). The algorithm is based on the classical primal-dual algorithm.The algorithm also gives rise to a transportation plan from one image to the other and we also show how this transportation plan can be used for interpolation and possibly also for compression and discrimination.     

M2+-树:一种支持医学病例多度量空间检索的高效索引

由于从病例库中进行病例的相似性检索关系到能否提供给医生充分且正确的候选病例,因此如何高效、准确地实现影像病例的相似性检索是学术界和医学界的研究热点之一.迄今为止,很多文献提出了用于提高查询精度的检索策略,但涉及检索效率的文章还为之甚少.基于此,提出了一种融多种度量空间相似性计算于一体的M2+-树高维索引技术.该索引将病例中的文本和影像合成一个高维多特征向量,该向量在度量空间上将数据空间划分成若干子空间,并借助关键向量对划分后的数据子空间再进行向量空间上的二次划分.关键向量的无重叠划分和三角不等式过滤原理可以加快病例的检索速度.总之,在度量和向量空间上的两次数据划分使得M2+-索引树大大减少了待查询病例与数据库病例间的不必要相似性计算的次数,从而加快了相似性病例的检索速度.实验结果表明,M2+-树的性能优于典型的度量空间多特征索引代表M2-树的性能.     

间隙度量与跟踪系统中的鲁棒控制器设计

为定量研究鲁棒控制器允许对象有尽可能大的不确定性.在引入间隙度量的基础上,定义了跟踪系统鲁棒控制器的鲁棒边界,并对某跟踪系统设计了基于间隙度量的鲁棒控制器.该控制器能兼顾对象的不确定性与控制器的不确定性.由仿真结果可以看出,与普通的PID控制器相比,具有较大鲁棒边界的鲁棒控制器不仅具有较强的干扰抑制能力.而且能够在模型加性不确定性存在的情况下具有很好的跟踪性能.     

A bounded distance metric for comparing tree structure

Comparing tree-structured data for structural similarity is a recurring theme and one on which much effort has been spent. Most approaches so far are grounded, implicitly or explicitly, in algorithmic information theory, being approximations to an information distance derived from Kolmogorov complexity. In this paper we propose a novel complexity metric, also grounded in information theory, but calculated via Shannon's entropy equations. This is used to formulate a directly and efficiently computable metric for the structural difference between unordered trees. The paper explains the derivation of the metric in terms of information theory, and proves the essential property that it is a distance metric. The property of boundedness means that the metric can be used in contexts such as clustering, where second-order comparisons are required. The distance metric property means that the metric can be used in the context of similarity search and metric spaces in general, allowing trees to be indexed and stored within this domain. We are not aware of any other tree similarity metric with these properties.