格子波尔兹曼方法的医学图像同步去噪增强算法

格子波尔兹曼的理论基础为分子动理学和统计力学,具有算法简单高效,执行速度快,易于并行处理等优点,近年来成功地应用于图像处理领域.本文针对医学图像反差较低且包含噪声污染的特点,通过设计分段线性拉伸函数作为格子波尔兹曼方程的外力项,实现医学图像的同步去噪和反差增强.同时构建基于TV下降流的PDE反差增强模型与本文的方法进行对比,实验表明,本文的方法处理效果优于TV下降流对应的效果.     

Parallel Lattice Boltzmann Method with Blocked Partitioning

This paper presents and discusses a blocked parallel implementation of bi- and three-dimensional versions of the Lattice Boltzmann Method. This method is used to represent and simulate fluid flows following a mesoscopic approach. Most traditional parallel implementations use simple data distribution strategies to parallelize the operations on the regular fluid data set. However, it is well known that block partitioning is usually better. Such a parallel implementation is discussed and its communication cost is established. Fluid flows simulations crossing a cavity have also been used as a real-world case study to evaluate our implementation. The presented results with our blocked implementation achieve a performance up to 31% better than non-blocked versions, for some data distributions. Thus, this work shows that blocked, parallel implementations can be efficiently used to reduce the parallel execution time of the method.     

LBM在流体渗透率计算中的应用与优化

格子Boltzmann方法广泛应用于流体计算领域,针对其应用中计算规模过大的问题提出了一种粒子空间压缩算法,改善了格子Boltzmann并行算法空间消耗过大的问题;合并了迁移与碰撞过程,减少了内存的I/O次数,加快了计算速度;该算法拥有良好的可扩展性。实验结果表明,优化后的程序空间消耗明显减少,性能提高50%以上。     

Optimization Strategies for the Entropic Lattice Boltzmann Method

The entropic formulation of the lattice Boltzmann method (LBM) features enhanced numerical stability due to its compliance with the Boltzmann H-theorem. This stability comes at the price of some computational overhead, associated with the need of adjusting the local relaxation time of the standard LBM in such a way as to secure compliance with the H-theorem. In this paper, we discuss a number of possible optimization strategies to reduce the computational overhead of entropic LBMs.     

基于Lattice Boltzmann模型的液-液混合流模拟

引入了一种二元Lattice Boltzmann Model（LBM）,实现丁两种液体组成的混合流的模拟．不同于其它的类似模型,它区分考虑了流体的粘性和扩散特性,可以很容易地模拟各种互溶或者不互溶的混合流现象．此外,由于LBM的运算大都是线性的局部运算,这使得它很容易在可编程图形处理器（Graphics Process Unit,GPU）上进行加速,从而进行实时模拟．给出了若干二元混合流的模拟结果．     

后台阶流动的非均匀格子Boltzmann方法模拟

使用非均匀格子Boltzmann方法对后台阶流动进行了数值模拟.将流体流动区域划分为不同的子区域:对于每个子区域内部,分布函数使用均匀网格计算;对于区域边界,分布函数采用嵌套网格方法进行处理.数值计算结果与其它实验、数值结果相吻合.     

A New Multiple-relaxation-time Lattice Boltzmann Method for Natural Convection

This article is devoted to the study of multiple-relaxation-time (MRT) lattice Boltzmann method with eight-by-eight collision matrix for natural convection flow. In the velocity space, eight speed directions are used and the corresponding incompressible multiple-relaxation-time model with force term is presented. D2Q4 model is for temperature field. The coupled double distribution functions (DDF) overcome artificial compressible effect corresponding to the standard MRT model. The simulations of natural convection flows with Pr=0.71 for air and Ra=10³–10⁹ are carried out and excellent agreements are obtained to demonstrate the numerical accuracy and stability of the proposed model.     

基于LBM算法的大规模人群疏散研究

人群聚集活动越来越多地出现在各种公共场合,而由此引发的各种安全问题也逐渐受到社会各界的重视.在重要建筑和人群密集场所突发重大恐怖事件时,将人员在较短的时间内紧急疏散至安全地带是恐怖事件紧急救援的首要任务.文中提出一种基于LBM的高效的人群疏散算法及模型,该模型将Lattice Boltzmann Method的D2Q9模型应用于人流疏散的模拟过程.该模f的着重点在于研究人群的疏散过程,从人流整体的运动特点考虑疏散问题.针对大规模人群紧急疏散的模拟,可比较精确地预测人流紧急疏散的过程与现象.     

一种不良图片快速过滤方法

本文提出了一种不良图片快速过滤的方法,该方法通过建立不同颜色空间上的混合肤色模型,用Adaboost算法进行人脸的准确定位,分析典型的肤色区域的几何特征等技术,能实现对包含不良信息图片的快速而准确的检测.实验结果显示,该方法对不良图片的检测平均可以达到80%以上的准确率,对于能够检测出人脸区域的图片,准确率可以达到98%.     

彩色图像混合滤波算法

提出了一种彩色图像混合滤波方法.该方法先判断像素的类型,然后用矢量加权均值滤波抑制高斯噪声,用多窗口矢量中值滤波抑制脉冲噪声.该算法能够自适应地选择滤波方法,提高了滤波性能.     

基于内容的特定图像过滤方法

针对互联网中色情图像传播愈来愈严重的现象,在充分分析色情图像的特征的基础上,提出基于计算机视觉与模式识别的图像过滤方法,该方法将人体肤色模型、面部模型以及图像轮廓、面积等多项图像特征识别技术相结合,实现网络色情图像识别过滤。该过滤方法,能够有效地过滤掉网络色情图像,实验结果表明,该方法能够达到80％以上的准确率,具有较高的实用性和应用价值。     

基于LBM的自由表面流体真实感绘制

针对自由表面流体的模拟,提出一种基于Lattice Boltzmann(LBM)的高效建模和绘制的方法.首先基于浅水方程的LBM模型进行流体建模及表面高度场计算,并提出一种基于Marching Cubes和自由表面算法结合的方法来抽取流体的表面,随后采用考虑移动障碍物的外力叠加机制和自适应加密算法来进行流体交互及表面的网格重构.最后采用硬件加速技术实现了不同自由表面流体的绘制,如溪流、水池浅水流、洪水水淹等真实感效果.     

图像插值的自适应邻域滤波方法

图像待插像素用其邻域内连续方向上像素的Tyler展开可以得到较好的近似,但是沿着灰度连续和不连续方向(跨越图像边缘方向)的Tyler展开近似的线性平均会增加图像边缘宽度,引起图像边缘的视觉模糊以及产生锯齿边缘。为此,通过一个与灰度距离相关的权函数,自适应选择待插像素邻域内的Tyler展开,提出一个邻域滤波图像插值方法。与待插像素位于图像边缘同侧的像素权函数的值较大,像素的Tyler展开被选择为待插像素的近似;反之则权函数的值较小,像素的Tyler展开不用作待插像素的Tyler展开。实验结果表明,该方法避免了跨越图像边缘的Tyler展开的线性平均,可减小插值图像边缘的宽度,增加边缘斜坡坡度,从而获得清晰的插值图像边缘。     

纹理图象的多信道滤波分割方法

纹理图象的分割分类方法是目前图象处理和机器视觉研究中的一个前沿课题，传统方法大多基于形态结构和统计描述，与人类视觉机理相脱节，无法进一步提高分精度。本文介绍了近几年来兴起的一类全新的方法，即基于空间／空间频率（ｓ／ｓｆ）平面的多信道滤波法，这类方法与人类视觉机理很好的吻合，对于人工纹理和自然纹理都能获很好的分割效果。     

多层次特定类型图像过滤方法

随着网络和多媒体技术的发展,互联网中色情图象的传播愈演愈烈,为有效杜绝此现象,提出了一种基于计算机视觉和模式识别的色情图像过滤方法－多层次特定类型图像过滤法,该方法通过研究人类视觉对图像的分析机理,提出综合肤色模型检验、支持向量机（SVM）分类和最近 邻近方法校验 的多层次图像处理框架,实验结果表明,该方法能够达到85％以上的准确率。     

Lattice Boltzmann simulations of drop dynamics

We present a free energy lattice Boltzmann approach to modelling the dynamics of liquid drops on chemically patterned substrates. We start by describing a choice of free energy that reproduces the bulk behaviour of a liquid–gas system together with the varying contact angles on surfaces with chemical patterning. After showing how the formulation of the free energy fits in to the framework of lattice Boltzmann simulations, numerical results are presented to highlight the applicability of the approach.     

A Diffusion Filtering Method for Image Processing

A new diffusion method for image filtering based on incorporating an integral of image intensity over a point neighborhood into the diffusion coefficient is suggested. The efficiency of the method is illustrated on a number of test images.     

ICA的一种新应用：图像滤波

利用独立分量分析的特性,提出一种新颖的图像滤波方法。该方法将原始图像和整体向右平移一列的图像看作ICA观测信号,对这2幅强相关性图像组成的矩阵信息进行ICA处理,可以得到一幅锐化图像,表明ICA在这种情况下具有高通滤波的性能。计算机仿真实验结果说明,该方法提取出的锐化图像具有层次感强、定位精度高等特点,是一种有效的滤波方法。     

Single bubble rising dynamics for moderate Reynolds number using Lattice Boltzmann Method

Dynamics of a single rising gas bubble is studied using a Lattice Boltzmann Method (LBM) based on the Cahn-Hilliard diffuse interface approach. The bubble rises due to gravitational force. However, deformation and velocity of the bubble depend on the balance of other forces produced by surface tension, inertia, and viscosity. Depending on the primary forces acting on the system, bubble dynamics can be classified into different regimes. These regimes are achieved computationally by systematically changing the values of Morton number (Mo) and Bond number (Bo) within the following ranges (1×10^-5<Mo<3×10⁴) and (1<Bo<1×10³). Terminal shape and Reynolds number (Re) are interactive quantities that depend on size of bubble, surface tension, viscosity, and density of surrounding fluid. Accurate simulation of terminal shape and Re for each regime could be satisfactorily predicted and simulated, since they are also functions of Mo and Bo. Results are compared with previous experimental results.     

Numerical Method Based on the Lattice Boltzmann Model for the Kuramoto-Sivashinsky Equation

In this paper, we proposed a lattice Boltzmann model based on the higher-order moment method for the Kuramoto-Sivashinsky equation. A series of partial differential equations obtained by using multi-scale technique and Chapman-Enskog expansion. According to Hirt’s heuristic stability theory, the stability of the scheme can be controlled by modulating some special moments to design the fifth-order dispersion term and the sixth-order dissipation term. As results, the Kuramoto-Sivashinsky equation is recovered with higher-order truncation error. The numerical examples show the higher-order moment method can be used to raise the accuracy of the truncation error of the lattice Boltzmann scheme for the Kuramoto-Sivashinsky equation.