基于特征尺度及多尺度分解的纹理分割

在计算机视觉领域,纹理识别及分割是图像低层次处理中一个重要的问题。文章依据视觉信息处理的生理学理论,提出了一种新的纹理分析算法。该算法通过搜寻纹理图像的特征尺度并配合连续整数尺度滤波上的特征尺度统计对纹理进行分层分析。并基于特征尺度上的分层分析提出一种纹理分割算法。实验结果证明了该算法的有效性。     

Multiscale traveling: crossing the boundary between space and scale

Adding multiscale interaction capabilities to 3D virtual environments may permit work with huge virtual worlds that might otherwise be too large to manage. Multiscale technology has shown potential to support user interactions. This paper reports an experimental study of two multiscale traveling techniques. Our results show that while allowing a flexible control on travel speed and accuracy is beneficial, directly traversing the space-scale could be a challenge for users, probably due to difficulties in perceiving scalable virtual space and executing scaling operations. The results suggest that more research is needed to improve the understanding of the coupling of space and scale in multiscale user interface and to harness the full potentials of multiscale traveling techniques.     

Multiscale fuzzy system identification

One of the biggest challenges in constructing empirical models is the presence of measurement errors in the data. These errors (or noise) can have a drastic effect on the accuracy and prediction of estimated models, and thus need to be removed for improved models accuracy. Multiscale representation of data has shown great noise-removal ability when used in data filtering. In this paper, this advantage of multiscale representation is exploited to improve the accuracy of the nonlinear Takagi–Sugeno (TS) fuzzy models by developing a multiscale fuzzy (MSF) system identification algorithm. The developed algorithm relies on constructing multiple TS fuzzy models at multiple scales using the scaled signal approximations of the input–output data, and then selecting the optimum multiscale model that maximizes the signal-to-noise ratio of the model prediction. The developed algorithm is shown to outperform the time domain fuzzy model, NARMAX model, and fuzzy model estimated from pre-filtered data using an Exponentially weighted Moving Average (EWMA) filter through a simulated shell and tube heat exchanger modeling example. The reason for this improvement is that the developed MSF modeling algorithm improves the model accuracy by integrating modeling and data filtering using a filter bank, from which the optimum filter (for modeling purposes) is selected.     

Numerical simulations of particle migration in a viscoelastic fluid subjected to Poiseuille flow

In this work we present 2D numerical simulations on the migration of a particle suspended in a viscoelastic fluid under Poiseuille flow. A Giesekus model is chosen as constitutive equation of the suspending liquid. In order to study the sole effect of the fluid viscoelasticity, both fluid and particle inertia are neglected.The governing equations are solved through the finite element method with proper stabilization techniques to get convergent solutions at relatively large flow rates. An Arbitrary Lagrangian–Eulerian (ALE) formulation is adopted to manage the particle motion. The mesh grid is moved along the flow so as to limit particle motion only in the gradient direction to substantially reduce mesh distortion and remeshing.Viscoelasticity of the suspending fluid induces particle cross-streamline migration. Both large Deborah number and shear thinning speed up the migration velocity. When the particle is small compared to the gap (small confinement), the particle migrates towards the channel centerline or the wall depending on its initial position. Above a critical confinement (large particles), the channel centerline is no longer attracting, and the particle is predicted to migrate towards the closest wall when its initial position is not on the channel centerline. As the particle approaches the wall, the translational velocity in the flow direction is found to become equal to the linear velocity corresponding to the rolling motion over the wall without slip.     

基于多尺度谱特征的SAR图像分割

该文结合多尺度技术与谱分析方法,提出了基于多尺度谱特征的图像分割方法,并将之用于SAR图像分割。该方法在多尺度框架内,提取每个像素在不同尺度下的局部谱特征(AR模型参数),并组合各尺度的谱特征为一多尺度谱特征向量,作为该像素的分类特征,利用一基于二元假设检验的分类器对该像素分类。与单一尺度的谱特征分割方法相比,多尺度谱特征分割保留了算法简单的优点的同时,在小窗口情况下,仍能给出较平滑的分割结果,从而减小了计算复杂度。     

运动图像多尺度分割中的参数估计方法

将一种新颖的多尺度分割方法引入了多分辨率运动图像分割框架中 ,从而形成了一种新的多尺度运动图像分割方法。该方法主要由多尺度随机场和序贯极大后验估计两个基本概念组成 ,主要是解决多尺度图像分割中仍然存在的计算复杂度问题 ,并可以通过上下文图像模型达到更加准确的图像分割。     

Multiscale computation for bioartificial soft tissues with complex geometries

The mechanical function of soft collagenous tissues is inherently multiscale, with the tissue dimension being in the centimeter length scale and the underlying collagen network being in the micrometer length scale. This paper uses a volume averaging multiscale model to predict the collagen gel mechanics. The model is simulated using a multiscale component toolkit that is capable of dealing with any 3D geometries. Each scale in the multiscale model is treated as an independent component that exchanges the deformation and average stress information through a scale-linking operator. An arterial bifurcation was simulated using the multiscale model, and the results demonstrated that the mechanical response of the soft tissues is strongly sensitive to the network orientation and fiber-to-fiber interactions.     

Multiscale recognition of legume varieties based on leaf venation images

In this work we propose an automatic low cost procedure aimed at classifying legume species and varieties based exclusively on the characterization and analysis of the leaf venation network. The identification of leaf venation patterns which are characteristic for each species or variety is not an easy task since in some situations (specially for cultivars from the same species) the vein differences are visually indistinguishable for humans. The proposed procedure takes as input leaf images acquired using a standard scanner, processes the images in order to segment the veins at different scales, and measures different traits on them. We use these features in combination with modern automatic classifiers and feature selection techniques in order to perform recognition. The process was initially applied to recognize three different legumes in order to evaluate the improvements over previous works in the literature, and then it was employed to distinguish three diverse soybean cultivars. The results show the improvements achieved by the usage of the multiscale features. The cultivar recognition is a more challenging problem, since the experts cannot distinguish evident differences in plain sight. However, we achieve acceptable classification results. We also analyze the feature relevance and identify, for each classifier, a small set of distinctive traits to differentiate the species and varieties.     

Mesh sensitivity in peridynamic simulations

In this work, we investigate the suitability of several meshing strategies for use with a common peridynamics solution scheme. First, we use a manufactured solution to quantify the influence of different meshes on the accuracy and conditioning of a nonlocal boundary value problem in one and two dimensions. We explore convergence behavior, the effects of model parameters, and sensitivity to perturbations. We then apply the same meshing strategies to a three-dimensional impact simulation that employs the full peridynamic mechanical theory. We present a qualitative comparison of the fracture patterns that result, and suggest best practices for generating meshes that lead to efficient, high-quality numerical simulations of peridynamic models.     

Multiscale facial structure representation for face recognition under varying illumination

Facial structure of face image under lighting lies in multiscale space. In order to detect and eliminate illumination effect, a wavelet-based face recognition method is proposed in this paper. In this work, the effect of illuminations is effectively reduced by wavelet-based denoising techniques, and meanwhile the multiscale facial structure is generated. Among others, the proposed method has the following advantages: (1) it can be directly applied to single face image, without any prior information of 3D shape or light sources, nor many training samples; (2) due to the multiscale nature of wavelet transform, it has better edge-preserving ability in low frequency illumination fields; and (3) the parameter selection process is computationally feasible and fast. Experiments are carried out upon the Yale B and CMU PIE face databases, and the results demonstrate that the proposed method achieves satisfactory recognition rates under varying illumination conditions.     

Automated adaptive cardiovascular flow simulations

We present an automatic adaptive procedure to perform blood flow simulations in the cardiovascular system. The procedure allows the user to start with subject-specific data collected through clinical measurements, like magnetic resonance imaging (MRI) data, and evaluate physiological parameters of interest, like flow distribution, pressure variations, wall shear stress, in an automatic and efficient manner. The process involves construction of geometric models of blood vessels, specification of flow conditions and application of an adaptive flow solver. The latter is based on incompressible Navier–Stokes equations using adaptive spatial discretization (meshing) techniques. In this article, we demonstrate the method on a model of a human abdominal aorta of a normal subject with geometry and flow rates assimilated from MRI data. The results obtained show that boundary layer mesh adaptivity offers a better alternative leading to more accurate predictions, especially for key physiological quantities like wall shear stress.     

Exploring pseudo- and chaotic random Monte Carlo simulations

Computer simulations are an increasingly important area of geoscience research and development. At the core of stochastic or Monte Carlo simulations are the random number sequences that are assumed to be distributed with specific characteristics. Computer-generated random numbers, uniformly distributed on (0, 1), can be very different depending on the selection of pseudo-random number (PRN) or chaotic random number (CRN) generators. In the evaluation of some definite integrals, the resulting error variances can even be of different orders of magnitude. Furthermore, practical techniques for variance reduction such as importance sampling and stratified sampling can be applied in most Monte Carlo simulations and significantly improve the results. A comparative analysis of these strategies has been carried out for computational applications in planar and spatial contexts. Based on these experiments, and on some practical examples of geodetic direct and inverse problems, conclusions and recommendations concerning their performance and general applicability are included.     

Using query-driven simulations for querying outcomes of business processes

When decision makers want to know outcomes of business processes in their organizations, they often use simulations to do this. Traditionally, this is achieved by running simulations, collecting statistics and answering questions of interests to the decision makers based on the collected statistics. This paper describes an alternative approach to simulations, called the Query-Driven Simulations (QDS) approach, in which the user first asks the queries of interest and then, depending on the query asked, appropriate simulations are run to answer that query. The QDS approach empowers the end-user by making it easier, faster, and more reliable to ask ad-hoc questions about outcomes of business processes than what is done in traditional simulations. To substantiate this point, the paper describes the types of questions decision makers ask about outcomes of business processes and studies how easy it is to express these questions in terms of an SQL-like query language SimQL designed for Query-Driven Simulations. In addition, a case study of using the QDS approach in a manufacturing application developed by a major management consulting company is presented and the QDS approach is tested on that application.     

Non-reversible Monte Carlo simulations of spin models

Monte Carlo simulations are used to study simple systems where the underlying Markov chain satisfies the necessary condition of global balance but does not obey the more restrictive condition of detailed balance. Here, we show that non-reversible Markov chains can be set up that generate correct stationary distributions, but reduce or eliminate the diffusive motion in phase space typical of the usual Monte Carlo dynamics. Our approach is based on splitting the dynamics into a set of replicas with each replica representing a biased movement in reaction-coordinate space. This introduction of an additional bias in a given replica is compensated for by choosing an appropriate dynamics on the other replicas such as to ensure the validity of global balance. First, we apply this method to a mean-field Ising model, splitting the system into two replicas: one trying to increase magnetization and the other trying to decrease it. For this simple test system, our results show that the altered dynamics is able to reduce the dynamical critical exponent. Generalizations of this scheme to simulations of the Ising model in two dimensions are discussed.     

Numerical simulations of laser wakefield accelerators in optimal Lorentz frames

The simulation of laser wakefield accelerators with particle-in-cell codes in relativistic reference frames is described, with emphasis on the computational speed-ups, which may potentially exceed three orders of magnitude in comparison with laboratory frame configurations. The initialization of laboratory quantities in a relativistically moving frame is depicted, and the method for result comparison with the plasma rest frame is described. Benchmarks with laboratory frame simulations and experimental data where gains of ∼20 times were obtained are discussed, and potential numerical issues are analyzed. This method enables numerical simulations with shorter turnaround times required for parameter scanning, and for one-to-one three-dimensional experimental modeling of current and next generation laser wakefield experiments.     

Reconsidering simulations in science education at a distance: features of effective use

Abstract  This paper proposes a reconsideration of use of computer simulations in science education. We discuss three studies of the use of science simulations for undergraduate distance learning students. The first one, The Driven Pendulum simulation is a computer-based experiment on the behaviour of a pendulum. The second simulation, Evolve is concerned with natural selection in a hypothetical species of a flowering plant. The third simulation, The Double Slit Experiment deals with electron diffraction and students are provided with an experimental setup to investigate electron diffraction for double and single slit arrangements. We evaluated each simulation, with 30 students each for The Driven Pendulum and Evolve simulations and about 100 students for The Double Slit Experiment . From these evaluations we have developed a set of the features for the effective use of simulations in distance learning. The features include student support, multiple representations and tailorability.     

Multi-block Poisson grid generator for cascade simulations

High quality computational grids can greatly enhance the accuracy of turbine and compressor cascade simulations especially when time-dependent results are sought where vortical structures are convected through the computational domain. A technique for generating periodic structured grids for cascade simulations based on the Poisson equations is described. To allow for more complex geometries, the grid can be divided into individual zones or blocks. The grids are generated simultaneously in all blocks, assuring continuity of the grid lines and their slopes across the zonal boundaries. Simple geometric rules can be employed for enforcing orthogonality at block boundaries. The method results in grids with low grid distortion by allowing both, block boundaries and grid points on physical boundaries, to move freely. Results are presented for a linear turbine and a linear compressor cascade.     

Multiscale directional filter bank with applications to structured and random texture retrieval

In this paper, multiscale directional filter bank (MDFB) is investigated for texture characterization and retrieval. First, the problem of aliasing in decimated bandpass images on directional decomposition is addressed. MDFB is then designed to suppress the aliasing effect as well as to minimize the reduction in frequency resolution. Second, an entropy-based measure on energy signatures is proposed to classify structured and random textures. With the use of this measure for texture pre-classification, an optimized retrieval performance can be achieved by selecting the MDFB-based method for retrieving structured textures and a statistical or model-based method for retrieving random textures. In addition, a feature reduction scheme and a rotation-invariant conversion method are developed. The former is developed so as to find the most representative features while the latter is developed to provide a set of rotation-invariant features for texture characterization. Experimental works confirm that they are effective for texture retrieval.