目标边界点集的层次化描述及其形状检索应用

提出用于轮廓线形状和区域形状图像检索的形状描述方法,该方法将目标形状的边界（包括内边界）表示为一个无序的点集,沿各方向对点集的迭代分割,建立层次化的边界点集描述模型.通过对各层形状边界的分割比和分散度的几何特征度量,产生各层的形状特征描述,对它们进行组合,建立对目标形状的层次化描述.两个目标形状的差异性度量定义为它们的层次化描述子的L-1距离.该方法具有：（1）通用性.能够描述轮廓线形状和区域形状这两种不同类型的形状;（2）可扩展性.基于所提出的分层描述框架,可以将分割比和分散度这两种几何度量进行扩展,纳入更多其他几何特征度量,以进一步提高形状描述的精度;（3）多尺度描述特性.提出的分层的描述机制,使得描述子具有内在的由粗到细的形状表征能力;（4）较低的计算复杂性.由于仅仅计算目标图像的边界像素点,使得算法具有较高的计算效率.用MPEG-7 CE-2区域形状图像库和MPEG-7 CE-1轮廓线形状图像库这两个标准测试集对该方法进行评估,并与同类的其他形状描述方法进行比较,实验结果表明：提出的方法在综合考虑检索精确率、检索效率和一般应用能力等指标的情况下,其性能上要优于各种参与比较的方法.     

基于方向特征的二值商标图像检索方法

文中针对二值商标图像，提出一种基于方向特征的商标图像检索方法。用边界方向直方图表示目标的边界方向特征，图像的相似性度量采用直方图求交算法；用区域方向直方图表示目标的区域方向特征。图像的相似性程度用直方图的相关距离来度量。这两者结合起来得到的方向特征可以同时反映图像的整体形状和轮廓形状。实验表明，方向特征能够有效地描述图像的形状及空间分布信息，取得了令人满意的检索结果。     

基于曲率和熵矩阵特征的商标图像检索

针对二值商标图像的特点,提出了一种综合边界曲率特征和图像分块熵矩阵特征的检索算法。首先,根据微分几何中曲率的定义,计算图像形状边界上每一点的曲率,并统计得到曲率直方图作为边界特征。然后,在图像分块的基础上,计算每一分块子图像的信息熵,得到熵矩阵,求该矩阵的奇异值作为区域特征。最后,综合这两个特征进行检索。实验表明,边界和区域特征综合使用较之单一特征有着更好的检索效果,并具有较好的几何不变性。     

基于改进Fast MBD显著性检测和多特征融合匹配的靶纸区域快速检测算法

为了解决基于机器视觉的自动报靶系统快速、准确定位靶纸区域的问题,通过对靶纸图像的颜色和形状特性分析,提出一种基于改进Fast Minimum Barrier Distance显著性和多特征匹配的靶纸区域快速检测算法。该算法在原始Fast Minimum Barrier Distance显著区域提取算法的基础上,引入局部区域对比度先验和形状先验作为显著性区域提取的补充准则。同时,为了判断提取到的区域是否包含靶纸,再引入多特征匹配算法。首先,分别对图像边界连通先验、局部区域对比度先验和形状先验进行量化,形成距离图、对比度图和形状图,再结合三者分割出显著性目标区域,然后提取分割出的各目标区域的多种特征,并将其进行特征融合,最后利用1-范式将得到的目标特征与模板特征进行匹配,把匹配结果小于阈值的目标视为靶纸。在400张包含靶纸图像数据集上的实验结果显示了该算法的有效性。同时,在华为海思平台上,该算法处理速度能达到30帧/秒,足以证明该算法的实时性。     

一种基于Zernike不变矩和改进的Hough变换的图像检索方法

针对基于内容的图像检索为了解决不变矩只关心对象的区域特征,而忽视了图像边界的缺点,提出Zernike不变矩与改进的Hough变换相结合的方法。改进的Hough变换方法充分利用Hough变换空间的统计信息,对形状的仿射变换具有一定的鲁棒性、易并行处理和良好的几何解析性等优点。该算法有效地融合了图像形状的边缘和区域特征。实验结果表明,在图像检索的综合性能上优于Zernike方法以及基于改进的Hough变换的方法,检索速度快,准确率高。     

基于区域伪Zernike矩的商标图像检索

研究商标图像检索问题,针对在二值商标图像检索中要求准确、高效,因此确定出目标像素点的区域分布形式,解决商标形状特征,对平移、旋转、比例以及起始点等变化下特征不变性问题,提出了一种基于区域伪Zernike矩的提取图像形状特征的算法。算法根据图像目标区域的最小外接圆进行区域划分,提取区域的伪Zernike矩来描述图像形状特征进行商标检索。通过仿真实验证明,区域伪Zernike矩的提取图像形状特征的算法具有良好的平移,旋转及比例不变性,并用有较强的抗噪性。     

利用三维自适应分裂-合并的MRI图像分割算法设计

为了在医学图像分割中,发现均匀几何三维区域的复杂形状,以提高分割准确率,提出一种基于3D几何特征分裂-合并(ASM)的脑部MRI图像分割算法;首先构建简单平行六面体的12种3D区域分割策略,体积分割技术将整个体积划分为许多大的均匀三维几何区;然后,在体积内定义更多小的均匀区域,以便在随后的合并步骤中有更大的生存概率;最后,进行多级区域合并,合并阶段只涉及复杂ASM树的叶子,考虑灰度相似性和共同边界区的大小,将小的区域合并为大邻近区;相比其他几种MRI图像分割算法,提出的方法在分割过程对噪声具有鲁棒性,提高了分割性能和准确率;另外提出的方法不需要训练数据集。     

基于五官特征的人脸剪纸生成方法研究

人脸肖像剪纸应该重现生动的图像细节,为了实现这一目标,提出了一种基于五官特征与图像变形算法的两阶段人脸剪纸合成方法。收集艺术家的人脸剪纸创作,分割五官部位并提取各组件的几何特征,建立数字化五官剪纸数据库。对目标人脸图像进行剪纸合成：在第一阶段,标定目标人脸图像的特征点,分割其五官部位,并提取各部位的几何特征,之后分别计算目标人脸五官与剪纸数据库中各对应组件基于几何特征和形状上下文特征的相似性度量值;通过融合几何特征和形状上下文特征,选择匹配相似度较高的剪纸部位,拼接得到初步的人脸剪纸图。在第二阶段,采用薄板样条（Thin Plate Spline,TPS）变形算法对第一阶段合成的人脸剪纸图进行变形,得到最终的剪纸图像。通过多人视觉测评实验,结果表明运用该方法得到的人脸剪纸图能够达到较为满意的效果。     

新的连通域标记方法及其在医学图像中的应用*

为了提高诊断的准确率和效率,提出了一种新的连通域标记算法,同时对医学图像中感兴趣区域进行连通域标记和区域特征提取。该算法先对读入二值图进行边界提取,再对边界进行跟踪和标记,利用图像重构的方法对边界进行区域填充,并将属于同一连通域的边界进行归类,即重新排列标记号,最后对连通区域的形态特征进行提取。实验证明,该算法不但能正确标记任意复杂形状的连通域,运行速度较快,而且对连通区域进行了特征提取,现已应用到医学图像处理的多个方面,为下一步的图像处理奠定了更好的基础。     

基于边缘几何不变性的特征提取算法研究

针对异源图像中可见光与红外图像特征提取算法进行研究,提出了一种基于边缘几何不变性特征的提取算法。采用空域滤波及灰度处理的方法对背景噪声进行处理,使用C anny算子提取目标的边缘,利用二值形态学中腐蚀与膨胀两种基本运算对边缘细化填充。利用圆形模板匹配的方法提取边缘特征点,通过对有效特征点的筛选与组合形成不同的三角形区域特征,并计算这些三角形区域的几何特性。仿真实验结果表明,该方法可以有效提取异源图像的共有特征。     

二维几何特征自适应有限元网格生成(一)--几何特征识别

发现二维形体边界均匀离散点集Delaunay三角剖分所具有的4个性质．根据这4个性质所描述的Delaunay三角剖分和形体几何特征之间的关系提出几何特征自动识别方法,并建立网格自适应机制,实现三维形体几何特征和部分力学特性自适应有限元网格自动生成．     

Design of internal pins in injection mold CAD via the automatic recognition of undercut features

A methodology to deal with the automatic design of internal pins in injection mold CAD via the automatic recognition of undercut features is developed. The approach to automatically identifying the undercut features is first proposed. For the given parting directions, all the inner and outer undercut features are identified based on the topological relationship of geometrical entities. The outer edge loop, which represents the largest cross-section boundary along the given parting directions, is extracted and patched up. The surfaces of molding are then identified based on the classifications of their geometrical entities. To identify the deep inner undercuts in the molding, the projection of the main core, internal pins and their bounding boxes along the parting direction and the pin withdrawal direction are generated. Upon determination of whether the bounding boxes of any two internal pins and the main core projection have intersection area, the deep inner undercuts are located. The complete methodology is finally implemented and verified through case studies and the efficiency of the methodology in handling complex molded parts is thus illustrated.     

Sketching freeform meshes using graph rotation functions

We present a new approach for sketching free form meshes with topology consistency. Firstly, we interpret the given 2D curve to be the projection of the 3D curve with the minimum curvature. Then we adopt a topology-consistent strategy based on the graph rotation system, to trace the simple faces on the interconnecting 3D curves. With the face tracing algorithm, our system can identify the 3D surfaces automatically. After obtaining the boundary curves for the faces, we apply Delaunay triangulation on these faces. Finally, the shape of the triangle mesh that follows the 3D boundary curves is computed by using harmonic interpolation. Meanwhile our system provides real-time algorithms for both control curve generation and the subsequent surface optimization. With the incorporation of topological manipulation into geometrical modeling, we show that automatically generated models are both beneficial and feasible.     

曲面重构中点云数据的区域分割研究

在曲面重构中，由于实际的曲面模型往往含有多个曲面几何特征，即由多张曲面组成，如果对使用激光法测量的“点云”数据直接进行拟合，将会造成曲面模型的数学表示和拟合算法处理的难度加大，甚至无法用较简单的数学表达式描述曲面模型，因此针对该问题，提出了一种基于数据点曲率变化的区域分割方法，即先对每一条扫描线上的数据点求取曲率值，然后将其中曲率值变化较大的点提取出来作为边界点，当边界确定后，再将云点数据分割成多个区域，由于每个区域一般具有较简单的几何特征，因此可用简单的数学模型来描述，并可重构单张曲面。该算法不仅原理简单、易于理解和编程，而且能提高曲面模型重构效率。     

基于聚类和PEDCC-Loss的CNN分类器的分类性能提升

提出了一种基于PEDCC-Loss和聚类的方法来提升CNN分类器分类性能的算法。利用CNN以及PEDCC-Loss来对图像进行特征提取,然后用BIRCH聚类算法对每类图像的隐特征进行聚类,以获得更好、更逼真的非线性边界,最大程度地减少误分类的边界点。将网络最后一层的PEDCC权重作为每类图像的中心,并以聚类后的子簇的簇心作为分类的辅助判断依据进行图像分类。实验结果表明,该算法的分类准确率相比CNN有一定的提升。     

Boundary analysis and geometric completion for recognition of interacting machining features

Features are the basic elements which transform CAD data into instructions necessary for automatic generation of manufacturing process plans. In this paper, a hybrid of graph-based and hint-based techniques is proposed to automatically extract interacting features from solid models. The graph-based hints generated by this approach are in geometrical and topological compliance with their corresponding features. They indicate whether the feature is 2.5D, floorless or 3D. To reduce the product model complexity while extracting features, a method to remove fillets existing in the boundary of a 2.5D feature is also proposed. Finally, three geometric completion algorithms, namely, Base-Completion, Profile-Completion and 3D-volume generation algorithms are proposed to generate feature volumes. The base-completion and profile-completion algorithms generate maximal volumes for 2.5D features. The 3D volume generation algorithm extracts 3D portions of the part.     

A computer-aided system for malignancy risk assessment of nodules in thyroid US images based on boundary features

In this paper, a novel computer-based approach is proposed for malignancy risk assessment of thyroid nodules in ultrasound images. The proposed approach is based on boundary features and is motivated by the correlation which has been addressed in medical literature between nodule boundary irregularity and malignancy risk. In addition, local echogenicity variance is utilized so as to incorporate information associated with local echogenicity distribution within nodule boundary neighborhood. Such information is valuable for the discrimination of high-risk nodules with blurred boundaries from medium-risk nodules with regular boundaries. Analysis of variance is performed, indicating that each boundary feature under study provides statistically significant information for the discrimination of thyroid nodules in ultrasound images, in terms of malignancy risk. k-nearest neighbor and support vector machine classifiers are employed for the classification tasks, utilizing feature vectors derived from all combinations of features under study. The classification results are evaluated with the use of the receiver operating characteristic. It is derived that the proposed approach is capable of discriminating between medium-risk and high-risk nodules, obtaining an area under curve, which reaches 0.95.     

3D clipping algorithm for feature mapping across domains

This paper describes an algorithm based on 3D clipping for mapping feature models across domains. The problem is motivated by the need to identify feature models corresponding to different domains. Feature mapping (also referred to as feature conversion) involves obtaining a feature model in one domain given a feature model in another. This is in contrast to feature extraction which works from the boundary representation of the part. Most techniques for feature mapping have focused on obtaining negative feature models only. We propose an algorithm that can convert a feature model with mixed features (both positive and negative) to a feature model containing either only positive or only negative features.The input to the algorithm is a feature model in one domain. The algorithm for mapping this model to another feature model is based on classification of faces of features in the model and 3D clipping. 3D clipping refers to the splitting of a solid by a surface. The feature mapping process involves three major steps. In the first step, faces forming the features in the input model are classified with respect to one another. The spatial arrangement of faces is used next to derive the dependency relationship amongst features in the input model and a Feature Relationship Graph (FRG) is constructed. In the second step, using the FRG, features are clustered and interactions between features (if any) are resolved. In the final step, the 3D clipping algorithm is used to determine the volumes corresponding to the features in the target domain. These volumes are then classified to identify the features for obtaining the feature model in the target domain. Multiple feature sets (where possible) can be obtained by varying the sequence of faces used for clipping. Results of implementation are presented.     

Static analysis of cylindrical shells by using B spline functions

This paper is concerned with the application of cubic B₃ spline, quintic B₅ spline functions and eigen-functions which satisfy the boundary conditions to obtaining the approximate solution for the static analysis of a flat cylindrical shell. Unified computational schemes suited for various types of boundary conditions of cylindrical shell are presented in the paper. In comparison with the conventional finite element method and finite strip method, the main features of the present method are higher accuracy, less unknowns, easier programming and more economy in computing storage and time requirement. The numerical results presented are in good agreement with the exact solution.     

An adaptive process planning approach of rapid prototyping and manufacturing

This paper presents an adaptive approach to improve the process planning of Rapid Prototyping/Manufacturing (RP/M) for complex product models such as biomedical models. Non-Uniform Rational B-Spline (NURBS)-based curves were introduced to represent the boundary contours of the sliced layers in RP/M to maintain the geometrical accuracy of the original models. A mixed tool-path generation algorithm was then developed to generate contour tool-paths along the boundary and offset curves of each sliced layer to preserve geometrical accuracy, and zigzag tool-paths for the internal area of the layer to simplify computing processes and speed up fabrication. In addition, based on the developed build time and geometrical accuracy analysis models, adaptive algorithms were designed to generate an adaptive speed of the RP/M nozzle/print head for the contour tool-paths to address the geometrical characteristics of each layer, and to identify the best slope degree of the zigzag tool-paths towards achieving the minimum build time. Five case studies of complex biomedical models were used to verify and demonstrate the improved performance of the approach in terms of processing effectiveness and geometrical accuracy.