一种基于特征线的曲面网格分割方法

各类网格分割法将曲面网格进行分割后,各子网格区域之间的交界线便可以作为曲面网格的封闭特征线。相反,如果根据网格模型的几何、拓扑特征,确定了网格模型的封闭特征线后,网格曲面便被这些特征线分割开来。为此,从曲面网格封闭特征线的角度出发,提出一种基于特征线的曲面网格分割方法。实验验证了该方法的可行性和有效性。     

Surface mesh regeneration considering curvatures

This work describes an automatic algorithm for unstructured mesh regeneration on arbitrarily shaped three-dimensional surfaces. The arbitrary surface may be: a triangulated mesh, a set of points, or an analytical surface (such as a collection of NURBS patches). To be generic, the algorithm works directly in Cartesian coordinates, as opposed to generating the mesh in parametric space, which might not be available in all the cases. In addition, the algorithm requires the implementation of three generic functions that abstractly represent the supporting surface. The first, given a point location, returns the desired characteristic size of a triangular element at this position. The second method, given the current edge in the boundary-contraction algorithm, locates the ideal apex point that forms a triangle with this edge. And the third method, given a point in space and a projection direction, returns the closest point on the geometrical supporting surface. This work also describes the implementation of these three methods to re-mesh an existing triangulated mesh that might present regions of high curvature. In this implementation, the only information about the surface geometry is a set of triangles. In order to test the efficiency of the proposed algorithm of surface mesh generation and implementation of the three abstract methods, results of performance and quality of generated triangular element examples are presented.     

基于多尺度局部特征的图像分割模型

为了解决图像分割中灰度不均匀和初始轮廓敏感的问题,提出一种基于多尺度局部特征的图像分割模型.与传统局部邻域定义在方形区域不同,该模型采用圆形区域来获取更多的局部信息;考虑到局部区域灰度的变化程度不一,提出利用多尺度结构与均值滤波器相结合的方法获得多尺度局部灰度信息;通过转换灰度不均匀模型得到一个逼近真实信息的图像,并将其融合进局部高斯分布拟合(LGDF)模型,构造出基于多尺度局部特征的能量泛函.从理论分析和实验结果表明:由于多尺度结构弱化了灰度不均匀的影响,该模型既能快速、准确地分割灰度不均匀图像,又表现出对初始轮廓具有较强的鲁棒性.     

Patch layout from feature graphs

The structuring of surface meshes is a labor intensive task in reverse engineering. For example, in CAD, scanned triangle meshes must be divided into characteristic/uniform patches to enable conversion into high-level spline surfaces. Typical industrial techniques, like rolling ball blends, are very labor intensive.We provide a novel, robust and quick algorithm for the automatic generation of a patch layout based on a topology consistent feature graph. The graph separates the surface along feature lines into functional and geometric building blocks. Our algorithm then thickens the edges of the feature graph and forms new regions with low varying curvature. Further, these new regions-so-called fillets and node patches-will have highly smooth boundary curves, making the algorithm an ideal preprocessor for a subsequent spline fitting algorithm.     

Combinatorial manifold mesh reconstruction and optimization from unorganized points with arbitrary topology

In this paper, a novel approach is proposed to reliably reconstruct the geometric shape of a physically existing object based on unorganized point cloud sampled from its boundary surface. The proposed approach is composed of two steps. In the first step, triangle mesh structure is reconstructed as a continuous manifold surface by imposing explicit relationship among the discrete data points. For efficient reconstruction, a growing procedure is employed to build the 2-manifold directly without intermediate 3D representation. Local and global topological operations with ensured completeness and soundness are defined to incrementally construct the 2-manifold with arbitrary topology. In addition, a novel criterion is proposed to control the growing process for ensured geometric integrity and automatic boundary detection with a non-metric threshold. The reconstructed manifold surface captures the object topology with the built-in combinatorial structure and approximates the object geometry to the first order. In the second step, new methods are proposed to efficiently obtain reliable curvature estimation for both the object surface and the reconstructed mesh surface. The combinatorial structure of the triangle mesh is then optimized by changing its local topology to minimize the curvature difference between the two surfaces. The optimized triangle mesh achieves second order approximation to the object geometry and can serve as a basis for many applications including virtual reality, computer vision, and reverse engineering.     

A hybrid approach to feature segmentation of triangle meshes

Segmentation of a polygonal mesh is a method of breaking the mesh down into ‘meaningful’ connected subsets of meshes called regions or features. Several methods have been proposed in the past and they are either vertex based or edge based. The vertex method used here is based on the watershed segmentation scheme which appears prominently in the image segmentation literature and was later applied to the 3D segmentation problem [9] and [10]. Its main drawback is that it is a vertex based method and no hard boundaries (edges) are created for the features or regions. Edge based methods rely on the dihedral angle between polygon faces to determine if the common edge should be classified as a Feature Edge. However, this method results in many disconnected edges and thereby incomplete feature loops.We propose a hybrid method which takes advantage of both methods mentioned earlier and create regions with complete feature loops. Satisfactory results have been achieved for both CAD parts as well as other laser scanned objects such as bones and ceramic vessels.     

基于透视点和区块特征的多车粘连分割方法

针对交通视频车辆检测与跟踪中经常出现的车辆前景粘连问题,提出了一种利用透视点在图像轮廓上搜索车辆分割点并通过区块特征识别车辆的粘连车辆分割方法.根据路面场景中车道线信息提取出透视点和车道区域,结合混合高斯模型与形态学梯度轮廓算法提取出车辆前景团块.利用透视点原理从前往后顺序分割粘连的车辆前景团块.对分割开的待检定区域,利用车辆区块特征进行检验识别,修正错误分割,将粘连的多辆车逐一分割.实验结果表明,该方法对直线道路上多辆车粘连的分割有较好的准确度和适应性.     

基于排列组合熵和灰度特征的纹理分割

提出了一种基于排列组合熵和灰度特征的纹理分割方法。该方法将不同方向上的排列组合熵与灰度均值、灰度方差结合起来构成一个多维特征向量,利用模糊C均值聚类算法进行聚类实现纹理图像的分割。实验结果表明该方法对纹理分布均匀的图像有着良好的分割效果。在保持较高纹理分割精度的前提下,该方法能减小计算复杂度,并且具有较强的鲁棒性和抗噪声能力。     

Consistent segmentation of 3D models

This paper proposes a method to segment a set of models consistently. The method simultaneously segments models and creates correspondences between segments. First, a graph is constructed whose nodes represent the faces of every mesh, and whose edges connect adjacent faces within a mesh and corresponding faces in different meshes. Second, a consistent segmentation is created by clustering this graph, allowing for outlier segments that are not present in every mesh. The method is demonstrated for several classes of objects and used for two applications: symmetric segmentation and segmentation transfer.     

基于多尺度特征融合双U型皮肤病变分割算法

针对皮肤病分割问题中皮肤病变区域大小不一且形状各异问题,提出一种基于多尺度特征融合的双U型皮肤病分割算法.该算法由粗分U型网络和细分U型网络两部分组成.首先粗分U型网络编码部分采用预训练VGG-19模型对相关特征进行多尺度特征提取;在解码阶段利用改进注意力残差块将底层与高层信息进行有效的映射融合,得到初步的Mask;然后将初步生成的Mask与原图像聚合,并输入多路特征提取编码器中进行二次特征蒸馏;而细分U型网络解码器同时与粗分U型网络编码部分和细分U型网络的编码部分特征映射进行融合,保证网络可以聚合更多的有效特征;最后利用Focal Tversky损失函数进一步提升分割效果.实验表明,所提算法在ISBI2016数据集上实验分割精度为96.11％、敏感度为93.59％、特异性为97.10％、Dice系数为93.14％、Jaccard系数为87.17％,能够有效地分割皮肤病病变区域.     

MSTAR图像的矩特征分析与多阈值分割

图像分割是SAR图像自动目标识别应用中的基础性问题。定义并分析了MSTAR图像的矩特征,进而构造了多阈值处理策略,用于MSTAR图像的分割。首先研究了目标、阴影,以及背景区域统计特性,并确定了相应的数学模型描述,在此基础上给出了矩特征的定义,并分析了其基本特性。通过由图像空间到矩特征空间的转换,显著增强了目标区域与阴影、背景区域的差异性,因此通过构造不同的阈值化规则,实现了MSTAR图像中目标、阴影和背景区域的分割。对MSTAR图像的处理结果表明,与恒虚警率（CFAR）、最大类间方差（OTSU）、模糊C均值（FCM）和马尔可夫随机场（MRF）等典型分割算法相比,本文方法不需进行噪声抑制,但在分割效果和鲁棒性等方面性能更好,同时,对多尺度、多目标MSTAR图像的分割也显示出良好的适应性。     

多尺度特征融合工件目标语义分割

目的 目标语义特征提取效果直接影响图像语义分割的精度,传统的单尺度特征提取方法对目标的语义分割精度较低,为此,提出一种基于多尺度特征融合的工件目标语义分割方法,利用卷积神经网络提取目标的多尺度局部特征语义信息,并将不同尺度的语义信息进行像素融合,使神经网络充分捕获图像中的上下文信息,获得更好的特征表示,有效实现工件目标的语义分割。方法 使用常用的多类工件图像定义视觉任务,利用残差网络模块获得目标的单尺度语义特征图,再结合本文提出的多尺度特征提取方式获得不同尺度的局部特征语义信息,通过信息融合获得目标分割图。使用上述方法经多次迭代训练后得到与视觉任务相关的工件目标分割模型,并对训练权重与超参数进行保存。结果 将本文方法和传统的单尺度特征提取方法做定性和定量的测试实验,结果表明,获得的分割网络模型对测试集中的目标都具有较精确的分割能力,与单尺度特征提取方法相比,本文方法的平均交并比mIOU（mean intersection over union）指标在验证集上训练精度提高了4.52%,在测试集上分割精度提高了4.84%。当测试样本中包含的目标种类较少且目标边缘清晰时,本文方法能够得到更精准的分割结果。结论 本文提出的语义分割方法,通过多尺度特征融合的方式增强了神经网络模型对目标特征的提取能力,使训练得到的分割网络模型比传统的单尺度特征提取方式在测试集上具有更优秀的性能,从而验证了所提出方法的有效性。     

基于融合相位特征的视网膜血管分割算法

针对相位一致性特征对血管中心检测不足问题,提出基于融合相位特征的眼底视网膜血管分割算法。首先,预处理原始的视网膜图像;然后,对图像中每个像素构造4D的特征向量（包括Hessian矩阵、Gabor变换、条带选择组合位移滤波响应（B-COSFIRE）滤波、相位特征）;最后,采用支持向量机（SVM）进行像素分类,实现眼底视网膜血管的分割。其中,相位特征是将分别提取的相位一致性特征与Hessian矩阵特征进行小波融合后得到的一种新的融合相位特征。该特征既保留了相位一致性特征良好的血管边缘信息,又克服了相位一致性特征对血管中心检测的不足。在用于血管提取的数字视网膜图像（DRIVE）数据库上测得基于融合相位特征的视网膜血管分割算法的平均准确率（Acc）为0.9574,平均受试者工作曲线面积（AUC）为0.9702;且在单一特征进行像素分类提取血管的实验中,与使用相位一致性特征相比,使用融合相位特征进行像素分类提取血管的Acc由0.9191提高到0.9478,AUC由0.9359提高到0.9578。实验结果表明,融合相位特征比相位一致性特征更适用于基于像素分类的眼底视网膜血管分割算法。     

基于语义分割动态特征点剔除的SLAM算法

针对动态物体容易干扰SLAM建图准确性的问题,提出了一种新的动态环境下的RGB-D SLAM框架,将深度学习中的神经网络与运动信息相结合。首先,算法使用Mask R-CNN网络检测可能生成动态对象掩模的潜在运动对象。其次,算法将光流方法和Mask R-CNN相结合进行全动态特征点的剔除。最后在TUM RGB-D数据集下的实验结果表明,该方法可以提高SLAM系统在动态环境下的位姿估计精度,比现有的ORB-SLAM2的表现效果更好。     

基于特征调节器和双路径引导的RGB-D室内语义分割

针对室内场景图像语义分割结果不精确、显著图粗糙的问题,提出一种基于多模态特征优化提取和双路径引导解码的网络架构（feature regulator and dual-path guidance,FG-Net）。具体来说,设计的特征调节器对每个阶段的多模态特征依次进行噪声过滤、重加权表示、差异性互补和交互融合,通过强化RGB和深度特征聚合,优化特征提取过程中的多模态特征表示。然后,在解码阶段引入特征交互融合后丰富的跨模态线索,进一步发挥多模态特征的优势。结合双路径协同引导结构,在解码阶段融合多尺度、多层次的特征信息,从而输出更细致的显著图。实验在公开数据集NYUD-v2和SUN RGB-D上进行,在主要评价指标mIoU上达到48.5 %,优于其他先进算法。结果表明,该算法实现了更精细的室内场景图像语义分割,表现出了较好的泛化性和鲁棒性。     

基于对抗学习和多尺度特征融合的前列腺MR图像分割

前列腺MR图像的自动分割已被广泛应用于前列腺癌的诊断和治疗过程中,然而,由于前列腺的形状变化显著且与相邻组织的对比度低,传统的分割方法仍存在精度低、速度慢等缺点.生成对抗网络GAN在计算机视觉任务中展示出了优越的性能,因此提出了一种使用对抗学习的概念来训练分割网络的方法,实现前列腺MR图像端到端的自动分割.模型框架主要由分割网络和判别网络构成,分割网络生成分割预测图,判别网络判断输入来自真实标签还是分割预测.同时,在分割网络中集成了感受野模块RFB来获取和融合深度特征的多尺度信息,提高特征的识别率和鲁棒性,以提升网络的分割性能.在PROMISE12数据集上的验证结果显示,该模型的DSC和HD分别为89.56％ 和7.65 mm.     

A mesh optimization algorithm based on neural networks

We have developed a mesh simplification method called GNG3D which is able to produce high quality approximations of polygonal models. This method consists of two distinct phases: an optimization phase and a reconstruction phase. The optimization phase is developed by applying an extension algorithm of the growing neural gas model, which constitutes an unsupervised incremental clustering algorithm. The primary goal of this phase is to obtain a simplified set of vertices representing the best approximation of the original 3D object. In the reconstruction phase we use the information provided by the optimization algorithm to reconstruct the faces obtaining the optimized mesh as a result. We study the model theoretically, analyzing its main components, and experimentally, using for this purpose some 3D objects with different topologies. To evaluate the quality of approximations produced by the method proposed in this paper, three existing error measurements are used. The ability of the model to establish the number of vertices of the final simplified mesh is demonstrated in the examples.     

时空特征融合网络的多目标跟踪与分割

目的 多目标跟踪与分割是计算机视觉领域一个重要的研究方向。现有方法多是借鉴多目标跟踪领域先检测然后进行跟踪与分割的思路,这类方法对重要特征信息的关注不足,难以处理目标遮挡等问题。为了解决上述问题,本文提出一种基于时空特征融合的多目标跟踪与分割模型,利用空间三坐标注意力模块和时间压缩自注意力模块选择出显著特征,以此达到优异的多目标跟踪与分割性能。方法 本文网络由2D编码器和3D解码器构成,首先将多幅连续帧图像输入到2D编码层,提取出不同分辨率的图像特征,然后从低分辨率的特征开始通过空间三坐标注意力模块得到重要的空间特征,通过时间压缩自注意力模块获得含有关键帧信息的时间特征,再将两者与原始特征融合,然后与较高分辨率的特征共同输入3D卷积层,反复聚合不同层次的特征,以此得到融合多次的既有关键时间信息又有重要空间信息的特征,最后得到跟踪和分割结果。结果 实验在YouTube-VIS(YouTube video instance segmentation)和KITTI MOTS(multi-object tracking and segmentation)两个数据集上进行定量评估。在YouTub...     

基于特征增强与最小误差分割的变化检测方法

以河北省石家庄市2003年和2004年的专题制图仪(TM)遥感影像为例,针对各波段光谱特征,提出了一种基于地物特征增强的变化检测方法.在两期影像上对各类地物采样并计算样本在不同波段的均值、标准差等特征量,以确定波段组合运算的加权系数,计算特征增强图像,实现两期影像中所指定地物类型的特征增强;计算两期特征增强影像的差异影像;使用最小误差分割法获取变化检测结果.通过对比实验可知:方法提取变化区域总体精度达到90%,相对于传统的基于主成分分析(PCA)的变化检测方法,具有较高的检测精度,较好的可行性与适应性.     

基于组注意力卷积和特征重建的轻量分割算法

针对卷积神经网络的庞大参数量和计算量难以应用于移动设备或嵌入式设备的问题,提出冗余特征重建模块(redundant feature reconstruction,RFR)和组注意力卷积模块(group attention convolution,GAC),RFR模块使用较少的参数量提取重要的固有特征,通过线性算子重建冗...