点云数据的凹陷体判定算法

体三维显示是三维显示领域重要研究方向,当处理带凹陷体的点云数据时,凹陷体区域无法正确判定导致对应区域的颜色无法正常显示,会对显示精度造成严重影响.为了解决点云中凹陷体判定问题,为点云数据中的凹陷体作规范化定义,并根据凹陷体的空间特点提出点云数据凹陷体判定(PCDD)算法.算法将点云体素化,之后根据凹陷体区域体素的属性特...     

基于决策树分类的体绘制加速技术

为了提高体绘制速度,提出一种新的算法。该算法通过决策树对体素分类,同时采用行程编码辅助模型存储决策树分类结果。在遍历体素模型时,只访问感兴趣的体素分类,而忽略那些空的和不感兴趣的体素分类,减少了体素的计算量。实验结果表明,此算法不仅保持了图像的绘制质量,而且明显提高了体绘制速度。     

基于数学形态学的染色体自动计数算法

通过研究染色体自动分析技术在评估儿童急性淋巴细胞白血病预后中的应用,提出了一种基于数学形态学的染色体自动计数算法.算法首先对输入的染色体图像进行了预处理,接着用标号法给图像中的每个对象赋予惟一的标号.为了消除图像中噪声对统计结果的影响,按照设定的规则将噪声对象删除.对于含有粘连染色体的对象,通过腐蚀运算计算出实际染色体个数,并根据结果修正标号值,修正结果即为图像中的杂色体数目.实验证明,该算法可以自动统计出图像中的染色体数目.     

基于动态自适应体素生长的肺部 CT图像3维分割算法

为了解决传统体素生长算法中容易产生分割泄漏问题以及左右肺分离的难题,提出了一种动态自适应的体素生长算法,用于对肺部CT图像进行3维分割.该算法采用了半侧肺预分割的方式,并针对分割区域的局部统计信息动态调整分割参数,有效地避免了分割泄漏,并在准确度和鲁棒性方面得到了很大的改进,而且能够应用于纯3维环境的自动分割.该算法通过对多组实验数据,包括健康人和不同类型肺病患者的肺部CT图像进行试验,取得了良好的效果,为肺部组织的定量计算以及临床应用提供了可靠的基础.     

基于相邻层间相似性和空体素跳跃的体绘制加速算法研究

Splatting是经典的基于物序的直接体绘制方法,运算数据量的多少制约着算法绘制图像的速度。为了进一步提升绘制速度,采用基于相邻层间相似性和空体素跳跃相结合的方法进行加速,在读取数据过程中对图片中的三维纹理数据进行筛选,并使用足迹表对筛选后的三维纹理数据进行二维投影,利用相邻层间相似性计算每一个点的灰度值,并根据灰度值将数据分类,算出对成像没有影响的空体素,跳过其绘制过程从而加速算法。实验结果显示,该算法能够在保证绘制图像质量的基础上,在一定程度上解决和改善Splatting算法数据的空间相关性和运算效率的问题。     

基于改进空体素跳跃法的光线投射算法

提出一种针对计算统一设备架构(CUDA)存储器的访存优化策略。在此基础上,给出适用于CUDA存储器的改进空体素跳跃法,以减少对空体素采样点的处理,加快基于Phong光照模型的光线投射算法的运行速度。实验结果表明,该算法能在保证图像质量的前提 下,提高图像绘制速度。     

基于二维纹理样本的方向场引导的体纹理合成

提出一种考虑方向场的体纹理合成新方法.在二维纹理样本中搜索邻域,采用优化的邻域匹配算法逐个合成体素;在匹配二维纹理样本像素邻域和体素邻域时,考虑各体素上的方向,生成各向异性体纹理,并允许采用各向异性的二维纹理样本获得满意的合成结果.方向场可由过程式方法生成,也可由三维医学数据处理而得.实验结果表明,从动物的弥散张量图像得到的方向场能有效地合成真实的动物肌肉纹理,为计算机图形学提供了一种新的肌肉建模手段.     

基于体素相似度的医学图像层间插值方法

断层图像层间插值是医学图像三维重建的一个重要环节,但现有插值算法易引起图像边界模糊或效率低下的不足。为此提出一种基于体素相似度的医学图像层间插值算法。该算法根据断层图像的体素相关性和组织本身特征信息计算其体素的相似度,并利用其相似度对插值点进行分类插值。实验结果表明,和已有的算法比较,新算法较好地提高了插值图像的质量,并且计算量大为减少。     

改进的点到三角网距离快捷算法

针对现有基于体素结构的点到三角网距离算法中存在的可能因未找到最近三角形所在体素而导致计算结果错误,以及较远点距离计算速度慢的问题,对其进行改进,提出准确快捷的基于体素结构的距离算法.首先将每个三角形进行点采样并求出采样点所在体素,得到三角网的体素化结果;对于每个待计算点,通过邻接体素与点距离变化规律寻找三角网包围盒内与该点距离小于某半径值的所有体素,逐渐扩大该半径值直至找到非空体素为止;以找到的体素中最近三角形到该点的距离作为初始距离,再令该半径值等于初始距离与一个适当微小量之和,即可找到与该点最近的三角形,得到准确的距离.实例测试结果表明,文中算法在一般情况下运算速度较快,与现有算法相比,较远点到三角网距离的计算时间明显缩短.     

基于体模型优化的虚拟试衣算法

针对目前虚拟试衣人体模型存在的精度和时效上的问题,提出了一种基于体素八叉树的优化的体模型的高效、真实感强的虚拟试衣算法。将空间划分为若干体素,与人体表面相交的体素构成人体表面模型,且体素采用八叉树结构存储。通过Kinect采集的人物的深度图像来计算八叉树中体素的符号距离函数(SDF);使用Marching cube融合图像彩色纹理并渲染出逼真的三维人体模型;提出了碰撞检测响应算法。实验结果表明,提出的方法增强了人物模型的真实感,八叉树结构有效缩短了碰撞检测时间,具有较强的实用价值。     

A fast display method for volumetric data

Presented is a fast display method for volumetric data sets, which involves a slicebased method for extracting potentially visible voxels to represent visible surfaces. For a given viewing direction, the number of visible voxels can be trimmed further by culling most of the voxels not visible from that direction. The entire 3D array of voxels is also present for invasive operations and direct access to interior structures. This approach has been integrated on a low-cost graphic engine as an interactive system for craniofacial surgical planning that is currently in clinical use.     

Optimizing dot graphics for volumetric displays

Discusses a dot-graphics plotting technique that allows a control system to read voxel data at a constant rate, despite the random-scan nature of the writing method. If we assume the data is arranged in consecutive slices, then the constrained 3D optimization problem reduces to a 2D one. In this article, we investigate and seek to optimize the efficiency with which the voxels within each slice are plotted, in order to reduce the memory requirements. Moreover, we consider the effect of changing the plotting order of the voxels within each slice to shorten the total beam path and thus the total time required to plot the voxels. This increases the total number of visible voxels displayed in the volume     

Shell rendering

A structure model for volume rendering, called a shell, is introduced. Roughly, a shell consists of a set of voxels in the vicinity of the structure boundary together with a number of attributes associated with the voxels in this set. By carefully choosing the attributes and storing the shell in a special data structure that allows random access to the voxels and their attributes, storage and computational requirements can be reduced drastically. Only the voxels that potentially contribute to the rendition actually enter into major computation. Instead of the commonly used ray-casting paradigm, voxel projection is used. This eliminates the need for render-time interpolation and further enhances the speed. By having one of the attributes as a boundary likelihood function that determines the most likely location of voxels in the shell to be on the structure boundary, surface-based measurements can be made. The shell concept, the data structure, the rendering and measurement algorithms, and examples drawn from medical imaging that illustrate these concepts are described     

一种基于抓取簇和碰撞体素的工业零件抓取姿态检测算法

针对工业上常见的散乱堆叠零件的抓取问题,提出一种基于抓取簇和碰撞体素的抓取姿态检测算法。所提出的抓取簇是定义在零件上的连续抓取姿态集合,解决了传统方法中因采用离散固定抓取点而导致可抓取点丢失、筛选效率低的问题。先对料箱和场景点云进行体素化;然后把包含料箱或点云的体素标记为碰撞体素,并把与碰撞体素相邻的体素标记为风险体素,从而建立体素化的碰撞模型;接下来,根据抓取簇的几何性质计算出候选的抓取姿态及其对应的抓取路径;最后,通过检测抓取路径所经过的体素类型来完成快速的碰撞检测,从而筛选出最优抓取姿态。基于所提算法搭建了完整的Bin-Picking系统,并对多种实际工业场景中常见的零件进行仿真实验和实际抓取实验,结果表明：该算法能够快速、准确地检测出安全的抓取姿态,实际抓取的平均成功率达92.2%,料箱平均清空率达87.2%,较传统方法有明显提升,且抓取过程均未发生碰撞,可满足实际工业应用的要求。     

一种用于时空体元编解码存储的低计算量优化方法

针对时空网格体对象的编解码占用存储空间大的问题,提出了一种用于时空体元编解码存储的低计算量优化方法。首先以十六叉树索引结构为基础,构建了时空网格体元编解码的数学模型,实现体元对象标识和时空位置索引,并借助3DGIS的自动编解码方法,实现了时空网格体元对象编解码存储表示的换算;其次,采用伽罗华有限域理论,构建了网格体元的二进制编码矩阵和存储的低计算量优化算法,实现了体元对象编解码存储过程中的优化计算;最后,以某矿山的矿床空间块体数据为例,对网格体元编解码模型、存储表示换算以及低计算量优化算法进了实际应用,并与八叉树索引结构的Morton码进行比较和分析,结果表明：该方法可有效降低30%的编解码存储计算量,提高了存储网格体元对象的时空效率。     

Seven-layer deep neural network based on sparse autoencoder for voxelwise detection of cerebral microbleed

In order to detect the cerebral microbleed (CMB) voxels within brain, we used susceptibility weighted imaging to scan the subjects. Then, we used undersampling to solve the accuracy paradox caused from the imbalanced data between CMB voxels and non-CMB voxels. we developed a seven-layer deep neural network (DNN), which includes one input layer, four sparse autoencoder layers, one softmax layer, and one output layer. Our simulation showed this method achieved a sensitivity of 95.13%, a specificity of 93.33%, and an accuracy of 94.23%. The result is better than three state-of-the-art approaches.     

结合超体素和区域增长的植物器官点云分割

点云分割是点云识别与建模的基础。为提高点云分割准确率和效率，提出一种结合超体素和区域增长的自适应分割算法。根据三维点云的空间位置和法向量信息，利用八叉树对点云进行初始分割得到超体素。选取超体素的中心体素组成一个新的重采样后的密度均匀点云，降低原始点云数据处理量，从而减少运算时间。建立重采样后点云数据的K-D树索引，根据其局部特征得到点云簇。最后将聚类结果返回到原始点云空间。分别选取植物三个物候期的激光扫描点云，对该方法的有效性进行验证。实验结果表明，该方法分割后点云与手工分割平均拟合度达到93.38%，高于其他同类方法，且算法效率得到明显提升。     

3D object recognition method with multiple feature extraction from LiDAR point clouds

During autonomous driving, fast and accurate object recognition supports environment perception for local path planning of unmanned ground vehicles. Feature extraction and object recognition from large-scale 3D point clouds incur massive computational and time costs. To implement fast environment perception, this paper proposes a 3D recognition system with multiple feature extraction from light detection and ranging point clouds modified by parallel computing. Effective object feature extraction is a necessary step prior to executing an object recognition procedure. In the proposed system, multiple geometry features of a point cloud that resides in corresponding voxels are computed concurrently. In addition, a scale filter is employed to convert feature vectors from uncertain count voxels to a normalized object feature matrix, which is convenient for object-recognizing classifiers. After generating the object feature matrices of all voxels, an initialized multilayer neural network (NN) model is trained offline through a large number of iterations. Using the trained NN model, real-time object recognition is realized using parallel computing technology to accelerate computation.

     

Association rule-based feature selection method for Alzheimer’s disease diagnosis

A fundamental challenge that remains unsolved in the neuroimage field is the small sample size problem. Feature selection and extraction, which are based on a limited training set, are likely to display poor generalization performance on new datasets. To address this challenge, a novel voxel selection method based on association rule (AR) mining is proposed for designing a computer aided diagnosis (CAD) system. The proposed method is tested as a tool for the early diagnosis of Alzheimer’s disease (AD). Discriminant brain areas are selected from a single photon emission computed tomography (SPECT) or positron emission tomography (PET) databases by means of an AR mining process. Simultaneously activated brain regions in control subjects that consist of the set of voxels defining the antecedents and consequents of the ARs are selected as input voxels for posterior dimensionality reduction. Feature extraction is defined by a subsequent reduction of the selected voxels using principal component analysis (PCA) or partial least squares (PLS) techniques while classification is performed by a support vector machine (SVM). The proposed method yields an accuracy up to 91.75% (with 89.29% sensitivity and 95.12% specificity) for SPECT and 90% (with 89.33% sensitivity and 90.67% specificity) for PET, thus improving recently developed methods for early diagnosis of AD.     

增强现实中人工标志识别方法研究

人工标志识别包括标志区域识别和标志图案识别,它是实现增强现实的重要步骤。通过对包含人工标志的图像进行二值化,采用连通域提取的算法实现了标志区域识别。对不同标志图案识别问题,提出了连通域数量判别法和模板匹配法。实验结果表明,所提方法能较好地识别人工标志,为实现虚实场景的实时融合打下了基础。