基于假高栓难和区域合并的视频对象分割

提出了基于假设检验和区域合并的视频对象分割算法.初始分割采用分水岭算法,接着根据颜色相似性进行区域合并,然后利用光流场估计和全局运动估计获得全局运动的残余误差,最后以各个区域的残余误差数据进行假设检验和小区域验证来确定运动区域,通过组合所有的运动区域即可分割出具有准确边缘的完整视频对象.对MPEG-4测试序列的实验结果表明了本算法具有良好的分割性能.     

基于假设检验和区域合并的视频对象分割

提出了基于假设检验和区域合并的视频对象分割算法。初始分割采用分水岭算法，接着根据颜色相似性进行区域合并，然后利用光流场估计和全局运动估计获得全局运动的残余误差，最后以各个区域的残余误差数据进行假设检验和小区域验证来确定运动区域，通过组合所有的运动区域即可分割出具有准确边缘的完整视频对象。对MPEG-4测试序列的实验结果表明了本算法具有良好的分割性能。     

一种交互式识别几何图形的简易方法

介绍了一种识别多笔画几何草图简单方法。它利用图形的暂时邻接关系和全局几何特征来识别一些简单的几何图形其中有实线和虚线类型,选择和删除手势。随着图形的旋转和尺度的变化这些几何特征（凸包、规则多边形的最小面积、周长和面积的数量比）是不变的。尽管多笔画的方法在选择合适的超时时间上存在问题,但这种方法可以得到令人满意的识别率。虽然在简单性和健壮性上更侧重于前者,但该方法已被证实适合交互式应用。     

一种交互式识别几何图形的简易方法

介绍了一种识别多笔画几何草图简单方法.它利用图形的暂时邻接关系和全局几何特征来识别一些简单的几何图形其中有实线和虚线类型,选择和删除手势.随着图形的旋转和尺度的变化这些几何特征(凸包、规则多边形的最小面积、周长和面积的数量比)是不变的.尽管多笔画的方法在选择合适的超时时间上存在问题,但这种方法可以得到令人满意的识别率.虽然在简单性和健壮性上更侧重于前者,但该方法已被证实适合交互式应用.     

一种新的基于相似度的目标跟踪算法

研究了视频序列中的目标跟踪问题,提出一种新的基于Chebyshev距离的相似度目标跟踪算法．在跟踪过程中,首先计算参考区域与目标区域间的相似度;然后确定在序列中与被跟踪区域相匹配的区域;最后利用统计检验的方法确定跟踪判断的正确性．该方法可以避免估计概率分布密度函数的不足．运用该算法在运动方式已知和未知的情况下分别与卡尔曼滤波和Camshift算法进行了对比实验,结果表明,所提出算法的跟踪效果有所提高．     

基于角点特征的形状识别

根据飞机形状角点之间的距离,定义了一种新的多维距离特征向量,不同飞机具有不同的多维距离向量.计算多维距离特征向量之间的相关系数,比较多维距离特征向量之间的相关性,能将每种飞机从飞机模型库中识别出来.实验表明,这种新特征向量具有较好的稳定性和区分性,算法识别率高、速度快,几乎达到实时效果.     

基于角点特征的形状识别

根据飞机形状角点之间的距离，定义了一种新的多维距离特征向量，不同飞机具有不同的多维距离向量。计算多维距离特征向量之间的相关系数，比较多维距离特征向量之间的相关性，能将每种飞机从飞机模型库中识别出来。实验表明，这种新特征向量具有较好的稳定性和区分性，算法识别率高、速度快，几乎达到实时效果。     

适用于遮挡问题的目标跟踪算法

提出一种基于网格模型的目标跟踪算法．该算法首先进行遮挡区域检测，然后进行网格结点的运动估计和网格更新过程完成目标的多帧跟踪．改进的遮挡区域检测算法有效地提高了检测准确度，从而确保遮挡区域的准确跟踪；网格结点的运动估计是通过特征窗口运动补偿匹配完成，可以有效地克服块效应．实验证明，该算法解决了二维运动估计时网格模型在遮挡区域存在的问题，并可以有效地进行目标准确跟踪．     

轮廓矩不变量及其在物体形状识别中的应用

为了有效地刻画物体的形状特征，在基于区域的Hu矩不变量的基础上，构造了一种基于物体轮廓曲线的新的矩不变量，即轮廓矩不变量。该不变量不仅独立于物体本身的颜色和灰度级，而且具有平移、旋转和尺度不变性，因此可将轮廓矩不变量应用于物体形状的识别，为了能快速地进行物体形状识别，还讨论了小波边缘检测和轮廓的获取问题及其算法。实验表明，基于这种轮廓矩的识别算法具有很好的识别率。     

A Hypothesis Testing Approach to Semifragile Watermark-Based Authentication

This paper studies the problem of achieving watermark semifragility in watermark-based authentication systems through a composite hypothesis testing approach. Embedding a semifragile watermark serves to distinguish legitimate distortions caused by signal-processing manipulations from illegitimate ones caused by malicious tampering. This leads us to consider authentication verification as a composite hypothesis testing problem with the watermark as side information. Based on the hypothesis testing model, we investigate effective embedding strategies to assist the watermark verifier to make correct decisions. Our results demonstrate that quantization-based watermarking is more appropriate than spread-spectrum-based methods to achieve the semifragility tradeoff between two error probabilities. This observation is confirmed by a case study of an additive Gaussian white noise channel with a Gaussian source using two figures of merit: 1) relative entropy of the two hypothesis distributions and 2) the receiver operating characteristic. Finally, we focus on common signal-processing distortions, such as JPEG compression and image filtering, and investigate the discrimination statistic and optimal decision regions to distinguish legitimate and illegitimate distortions. The results of this paper show that our approach provides insights for authentication watermarking and allows for better control of semifragility in specific applications.      

Using the Median Distance to Compare Object Shapes in Content-Based Image Retrieval

Turning Angles (TAs) representation is considered one of the most interesting methods for representing object shapes in content-based image retrieval systems. Nevertheless, the distance commonly used to measure the similarity between shapes represented by TAs, the Euclidean one, is generally too sensitive to small variations in shapes.In this paper we present a new distance between shapes represented by TA, namely the Median distance, specially devised to minimize the effects of small variations in shapes. Its analytical properties are discussed and experimental results are provided and compared with those obtained by applying traditional techniques based on Euclidean distance. The Median distance has been implemented in the Automatic Image Storage and Retrieval (AISR) system, which allows storage and content-based retrieval of 2D images.     

几何不变性及其在3D物体识别中的应用

三维物体识别是计算机视觉研究的重要内容之一，它要求从3D物体的2D图象中识别和定位物体．由于物体成像时会受到观察视角、摄像机参数的影响，因此使得同一物体在不同观察视角、不同摄像机参数等条件下所得到的图象存在差异．但由于几何不变性方法可以有效地消除这种差异带给3D物体识别的不利影响，所以，近20年来这种方法受到了广泛的关注和研究．为使人们了解该领域的研究现状，以对该领域的研究有所启发，首先讨论了基于几何不变性的3D物体识别方法的研究内容，包括研究的几何框架和其不变性以及几何不变性在3D物体识别中的主要应用；其次，总结性地评述了该领域的研究现状；最后，提出了研究的发展方向．     

Planar Numerical Signature Theory Applied to Object Recognition

Geometric invariants play a crucial role in the field of object recognition where the objects of interest are affected by a group of transformations. However, designing robust algorithms that are tolerant to noise and image occlusion remains an open problem. In particular, numerical signature-invariants in terms of joint invariants, as an approximation to the differential signature-invariants, suffer instability, bias, noise and indeterminacy in the resulting signatures. This paper addresses some of these issues in respect of planar signatures. To improve the stability in the Euclidean case, we replace Heron’s formula by the “accurate area” and then we demonstrate that the proposed algorithm is, not only numerically stable but is also, in terms of mean square error, a closer approximation (by at least a factor of three) compared with the original formulation of Calabi. To reduce noise in the resulting curves “the n-difference technique” and “the m-mean signature method” are introduced and we show that these methods are capable of minimizing noise by more than 90 %. The n-difference technique can also be applied to eliminate indeterminacy in the outputs. For the equiaffine case, we improve and extend the required formulation for the implementation of Signature theory for any planar meshes with a general position property. Moreover, we introduce a general formulation for the full conic sections to determine an equiaffine-invariant numerical approximation to the equiaffine arc length, measured along the given curve between any two points of the mesh. Finally, we demonstrate the discriminative power of the concept of discrete signature analysis for distinguishing normal and abnormal regions in the medical imaging domain.     

Geometric Heat Equation and Nonlinear Diffusion of Shapes and Images

Visual tasks often require a hierarchical representation of shapes and images in scales ranging from coarse to fine. A variety of linear and nonlinear smoothing techniques, such as Gaussian smoothing, anisotropic diffusion, regularization, etc., have been proposed, leading to scalespace representations. We propose ageometricsmoothing method based on local curvature for shapes and images. The deformation by curvature, or the geometric heat equation, is a special case of thereaction–diffusionframework proposed in [41]. For shapes, the approach is analogous to the classical heat equation smoothing, but with a renormalization by arc-length at each infinitesimal step. For images, the smoothing is similar to anisotropic diffusion in that, since the component of diffusion in the direction of the brightness gradient is nil, edge location is left intact. Curvature deformation smoothing for shape has a number of desirable properties: it preserves inclusion order, annihilates extrema and inflection points without creating new ones, decreases total curvature, satisfies the semigroup property allowing for local iterative computations, etc. Curvature deformation smoothing of an image is based on viewing it as a collection of iso-intensity level sets, each of which is smoothed by curvature. The reassembly of these smoothed level sets into a smoothed image follows a number of mathematical properties; it is shown that the extension from smoothing shapes to smoothing images is mathematically sound due to a number of recent results [21]. A generalization of these results [14] justifies the extension of the entireentropy scale spacefor shapes [42] to one for images, where each iso-intensity level curve is deformed by a combination of constant and curvature deformation. The scheme has been implemented and is illustrated for several medical, aerial, and range images.     

Mutation Testing Applied to Estelle Specifications

Many researchers have pursued the establishment of a low-cost, effective testing and validation strategy at the program level as well as at the specification level. Mutation Testing is an error-based approach, originally introduced for program testing, that provides testers a systematic way to evaluate how good a given test set is. Some studies have also investigated its use to generate test sets. In this article, the application of Mutation Testing for validating Estelle specifications is proposed. A mutant operator set for Estelle—one of the crucial points for effectively applying Mutation Testing—is defined, addressing: the validation of the behavior of the modules, the communication among modules and the architecture of the specification. In this scope, these operators can be taken as a fault model. Considering this context, a strategy for validating Estelle-based specification is proposed and exemplified using the Alternating-bit protocol.     

Differential and Numerically Invariant Signature Curves Applied to Object Recognition

We introduce a new paradigm, the differential invariant signature curve or manifold, for the invariant recognition of visual objects. A general theorem of É. Cartan implies that two curves are related by a group transformation if and only if their signature curves are identical. The important examples of the Euclidean and equi-affine groups are discussed in detail. Secondly, we show how a new approach to the numerical approximation of differential invariants, based on suitable combination of joint invariants of the underlying group action, allows one to numerically compute differential invariant signatures in a fully group-invariant manner. Applications to a variety of fundamental issues in vision, including detection of symmetries, visual tracking, and reconstruction of occlusions, are discussed.     

Geometric Properties of Bisector Surfaces

This paper studies algebraic and geometric properties of curve–curve, curve–surface, and surface–surface bisectors. The computation is in general difficult since the bisector is determined by solving a system of nonlinear equations. Geometric considerations will help us to determine several distinguished curve and surface pairs which possess elementary computable bisectors. Emphasis is on low-degree rational curves and surfaces, since they are of particular interest in surface modeling.     

分段模型在解码假设检验中的应用

本文主要研究了分段模型(以参数轨迹模型为例)在解码假设检验中的应用。分段模型与传统的HMM相比,具有更加精确的建模能力。多年来人们一直致力于研究它对语音识别性能的提高,而忽视了其它方面的应用。本文提出了分段模型校验的方法,对HMM的识别结果进行二次处理,克服了传统方法在不同句子间不具有可比性的缺点,简单而有效;在此基础上,为了满足系统的特殊要求,训练Fisher分类器,选择分段模型而非HMM的N-Best信息作为特征输入,验证了分段模型得分作为可信度指标时的优秀区分能力。实验结果表明,在第一类错误率为5%的情况下,最好的第二类错误率可以降到25.265%。这体现了系统良好的拒识性能。