A Topology Mapping Method for Feature Extraction of Irregular Curve Shape

ＡＴｏｐｏｌｏｇｙＭａｐｐｉｎｇＭｅｔｈｏｄｆｏｒＦｅａｔｕｒｅＥｘｔｒａｃｔｉｏｎｏｆＩｒｒｅｇｕｌａｒＣｕｒｖｅＳｈａｐｅＧＵＯＺｉｈａｉ；ＷＡＮＧＷｅｉ；ＪＩＡＧｕａｎｇ郭子海，王薇，贾光（Ｄｅｐｔ．ｏｆＣｏｍｐｕｔｅｒＳｃｉｅｎｃｅａｎｄＥｎ...

Quality Evaluation Model for Map Labeling

This paper discusses and sums up the basic criterions of guaranteeing the labeling quality and abstracts the four basic factors including the conflict for a label with a label, overlay for label with the features, position’s priority and the association for a label with its feature. By establishing the scoring system, a formalized four-factors quality evaluation model is constructed. Last, this paper introduces the experimental result of the quality evaluation model applied to the automatic map labeling system-MapLabel.

On Edge Irregular Reflexive Labeling of Categorical Product of Two Paths

Among the huge diversity of ideas that show up while studying graph theory,one that has obtained a lot of popularity is the concept of labelings of graphs.Graph labelings give valuable mathematical models for a wide scope of applications in high technologies(cryptography,astronomy,data security,various coding theory problems,communication networks,etc.).A labeling or a valuation of a graph is any mapping that sends a certain set of graph elements to a certain set of numbers subject to certain conditions.Graph labeling is a mapping of elements of the graph,i.e.,vertex and for edges to a set of numbers(usually positive integers),called labels.If the domain is the vertex-set or the edge-set,the labelings are called vertex labelings or edge labelings respectively.Similarly,if the domain is V(G)[E(G)],then the labeling is called total labeling.A reflexive edge irregular k-labeling of graph introduced by Tanna et al.:A total labeling of graph such that for any two different edges ab and a'b'of the graph their weights has wt_(x)(ab)=x(a)+x(ab)+x(b) and wt_(x)(a'b')=x(a')+x(a'b')+x(b') are distinct.The smallest value of k for which such labeling exist is called the reflexive edge strength of the graph and is denoted by res(G).In this paper we have found the exact value of the reflexive edge irregularity strength of the categorical product of two paths (P_(a)×P_(b))for any choice of a≥3 and b≥3.

Variable-coordinate forward modeling of irregular surface based on dual-variable grid

The mapping method is a forward-modeling method that transforms the irregular surface to horizontal by mapping the rectangular grid as curved; moreover, the wave field calculations move from the physical domain to the calculation domain. The mapping method deals with the irregular surface and the low-velocity layer underneath it using a fine grid. For the deeper high-velocity layers, the use of a fine grid causes local oversampling. In addition, when the irregular surface is transformed to horizontal, the flattened interface below the surface is transformed to curved, which produces inaccurate modeling results because of the presence of ladder-like burrs in the simulated seismic wave. Thus, we propose the mapping method based on the dual-variable finite-difference staggered grid. The proposed method uses different size grid spacings in different regions and locally variable time steps to match the size variability of grid spacings. Numerical examples suggest that the proposed method requires less memory storage capacity and improves the computational efficiency compared with forward modeling methods based on the conventional grid.

基于MRF-MAP和模糊质量图的相位展开算法

相位展开是磁共振成像技术应用中最关键的环节之一,可以为磁共振的某些重要临床应用提供精确的相位信息。然而,由于临床磁共振成像过程中,部分区域真实的相位存在急剧变化,同时伴有不同性态的噪声污染,导致相位展开时存在信息的高度不一致性。为了有效地解决上述难题,基于马尔可夫-最大后验(Markov Random Field& Maximum A Posterioi,MRF-MAP)模型,首次将相位展开看作计算机视觉中的标记问题,并结合磁共振相位数据的特点,设计出相位图的模糊质量图,完成相位展开的能量函数构建。针对能量函数的优化求解,采用高效的图割算法进行。

基于MapX的GIS二次开发中的相关问题的探讨

本文主要研究了基于MapX组件的地理信息系统的二次开发。由于MapX本身存在的某些缺陷,不能够以中文字段值作为专题字段生成专题的问题与不支持多字段同时标注的问题,提出了采用构建离线ADO重新绑定MapX的Layer对象的方法和利用MapX的Annotation对象动态生成标注的方法来解决这两类问题,并在陕西师范大学公用房管理系统中成功地进行了实现。

天地图电子地图配图中Maplex智能标注方法研究

Maplex是一款基于ArcGIS软件的智能化高级标注引擎,具有一系列特殊拓展模块工具,从而实现更多复杂设置。文章详细介绍了Maplex的主要功能模块,并以天地图中地名地址与兴趣点数据为例,探讨了利用Maplex智能标注引擎解决天地图电子地图配图中标注过密、互相压盖等技术难题,显著提高了天地图地图标注的质量和效率。

PLDMLT:Multi-Task Learning of Diabetic Retinopathy Using the Pixel-Level Labeled Fundus Images

In the field of medical images,pixel-level labels are time-consuming and expensive to acquire,while image-level labels are relatively easier to obtain.Therefore,it makes sense to learn more information(knowledge)from a small number of hard-to-get pixel-level annotated images to apply to different tasks to maximize their usefulness and save time and training costs.In this paper,using Pixel-Level Labeled Images forMulti-Task Learning(PLDMLT),we focus on grading the severity of fundus images for Diabetic Retinopathy(DR).This is because,for the segmentation task,there is a finely labeled mask,while the severity grading task is without classification labels.To this end,we propose a two-stage multi-label learning weakly supervised algorithm,which generates initial classification pseudo labels in the first stage and visualizes heat maps at all levels of severity using Grad-Cam to further provide medical interpretability for the classification task.A multitask model framework with U-net as the baseline is proposed in the second stage.A label update network is designed to alleviate the gradient balance between the classification and segmentation tasks.Extensive experimental results show that our PLDMLTmethod significantly outperforms other stateof-the-art methods in DR segmentation on two public datasets,achieving up to 98.897%segmentation accuracy.In addition,our method achieves comparable competitiveness with single-task fully supervised learning in the DR severity grading task.

Building of cognizing semantic map in large-scale semi-unknown environment

The quick response code based artificial labels are applied to provide semantic concepts and relations of surroundings that permit the understanding of complexity and limitations of semantic recognition and scene only with robot's vision.By imitating spatial cognizing mechanism of human,the robot constantly received the information of artificial labels at cognitive-guide points in a wide range of structured environment to achieve the perception of the environment and robot navigation.The immune network algorithm was used to form the environmental awareness mechanism with "distributed representation".The color recognition and SIFT feature matching algorithm were fused to achieve the memory and cognition of scenario tag.Then the cognition-guide-action based cognizing semantic map was built.Along with the continuously abundant map,the robot did no longer need to rely on the artificial label,and it could plan path and navigate freely.Experimental results show that the artificial label designed in this work can improve the cognitive ability of the robot,navigate the robot in the case of semi-unknown environment,and build the cognizing semantic map favorably.

A New Improved Symbol Mapper/8-Ary Constellation for BICM-ID

This contribution proposes a new combination symbol mapper/8-ary constellation, which is a joint optimization of an 8-ary signal constellation and its symbol mapping operation, to improve the performance of Bit Interleaved Coded Modulation with Iterative Decoding (BICM-ID). The basic idea was to use the so called (1,7) constellation (which is a capacitive efficient constellation) instead of the conventional 8-PSK constellation and to choose the most suitable mapping for it. A comparative study between the combinations most suitable mapping/(1,7) constellation and SSP mapping/conventional 8-PSK constellation has been carried out. Simulation results showed that the 1st combination significantly outperforms the 2nd combination and with only 4 iterations, it gives better performance than the 2nd combination with 8 iterations. A gain of 4 dB is given by iteration 4 of the 1st combination compared to iteration 8 of the 2nd combination at a BER level equal to 10-5, and it (iteration 4 of the 1st combination) can attain a BER equal to 10-7 for, only, a SNR = 5.6 dB.

Excel与MapGis结合绘制钻孔柱状图的初步应用

在Excel表中直接输入或运用函数计算显示出一定格式的、MapGis可识别的数据,实现施工现场编录数据与MapGis软件的数据对接,通过一次数据投影获得钻孔柱状图的数据标注,可极大提高成图过程中数据录入的效率,节省大量人力资源,特别是在一些全孔取样的工程中更具有现实意义.

基于约束DT的大规模动态地形的构建

大规模地形场景模型复杂、数据量大,动态实时构建效率低。论文研究了大规模地形的动态构建问题,提出了一种基于约束DT进行地形动态实时构网的方法。该方法将大规模的地形分割成若干小的地形块,将地形块的四个顶点添加到地形数据集中,对每个地形块单独构网,从而实现大规模地形三角网的快速构建。针对地形的动态变化,采用添加扩展约束的方式仅对扩展约束内部动态变化的地形进行重新构网,提高了地形三角网动态更新的效率。在进行地形纹理贴图时,对地形块进行分块处理,建立纹理数据与地形子块内三角网的映射关系实现纹理贴图,提高了渲染效率。实验结果表明,地形分块动态建模提高了地形动态变化后构网和绘制的效率。

结合双流形映射的不完备多标签学习

在多标签学习中,有效利用标签相关性可以提高分类性能。然而,由于人工标注标签的主观性和实际应用中标签语义的相似性,通常只能观察到不完备的标签空间,导致标签相关性的估计不准确,使得算法性能下降。针对该问题,提出一种结合双流形映射的不完备多标签学习(ML-DMM)算法。构造两种流形映射,一种是保留实例数据空间局部结构信息的特征流形映射,另一种是基于迭代学习得到的标签相关性的标签流形映射。首先通过拉普拉斯映射构造数据的低维流形,然后通过回归系数矩阵和标签相关性矩阵将初始特征空间和初始标签空间分别映射到该低维流形上,形成一种双流形映射结构来提升算法性能,最后利用迭代学习得到的回归系数矩阵进行多标签分类。在8个多标签数据集及3种标签缺失率情况下的对比实验结果表明,ML-DMM算法性能优于其他针对缺失标签的多标签分类算法。

高速铁路千米级主跨悬索桥线-桥垂向变形映射关系分析

为研究千米级悬索桥线-桥垂向变形映射关系,逐层分析有砟轨道层间传力特性,建立悬索桥垂向变形与轨面几何形态的映射解析模型,并对映射模型的准确性和有效性进行验证。研究结果表明:桥梁变形波长大于16 m时,线-桥垂向变形可视为完全映射,波长小于16 m时,线-桥位移有所差异;扣件和道床刚度主要影响波长16 m以下的线-桥垂向变形映射关系;悬索桥上线路平顺性表现出显著的温度敏感性,影响桥上线路垂向不平顺的主要因素依次为主缆、主塔、吊杆和主梁;墩塔沉降作用下桥上线路产生不平顺的主要范围集中在桥梁边跨和主塔附近;桥上线路特征波长主要为28、14、28、56 m,分别对应桥梁的横梁间距、节间间距、标准节间长度和4倍节间间距。

基于双HSIC和稀疏正则化的多标签特征选择

为了合理地利用多标签数据中的样本信息和标签信息,提高模型的分类性能,提出了基于双希尔伯特-施密特独立性准则(Hilbert-Schmidt independence criterion,HSIC)和稀疏正则化的多标签特征选择(DHSR)。该方法在线性映射的基础上引入双HSIC作为正则项,增强伪标签空间和特征空间之间的依赖关系,增强伪标签空间和真实标签空间之间的依赖关系。并使用L2,1范数作为稀疏正则项,以提高模型的泛化能力和减少模型的计算复杂度。最后,在多个经典多标签数据集上的对比实验结果验证了DHSR的有效性和优越性。

注意力缺陷多动障碍的神经影像学研究进展

注意力缺陷多动障碍(ADHD)在学龄期儿童中较为常见,常以与年龄不相符的注意力不集中、日常过度活动和情绪冲动为特征,虽智力正常或接近正常,但给患儿的日常生活以及学习带来较多困难。近年来,ADHD的神经影像学研究发展迅速,如定量磁敏感成像(QSM)、扩散磁共振(dMRI)、合成磁共振(SyMRI)、动脉自旋标记(ASL)、磁共振波谱(MRS)、静息态磁共振(rs-fMRI)等技术可以为ADHD疾病的诊断提供影像学标志物。本文旨在对ADHD的神经影像学研究进行综述,探究ADHD患者潜在的神经病理学机制,帮助临床医生进行早期诊断、及时干预,以及判断患儿的预后情况。

类别条件生成对抗网络的语音对抗样本生成方法

针对现有面向自动语音识别系统的对抗攻击方法难以捕捉不同语音尺度之间的相关性、导致攻击成功率低的问题,提出一种类别条件生成对抗网络的语音对抗样本生成方法。通过目标标签映射模块,将攻击目标标签转化为独热向量,作为条件输入到构建的类别条件生成对抗网络中,以此控制语音样本类别的生成。该类别条件生成对抗网络中的生成器,采用设计的NReSidual U-block网络模块与U-Net相融合,可以更好地学习不同时间尺度的语音特征,以及提升语音特征的表示能力,从而可以针对特定语音类别生成对抗样本;判别器采用卷积块和全连接层相结合的网络结构,将错误损失通过梯度反向传播至生成器,能有效保留语音信号的时序信息,并解决数据分布不稳定问题。在通用的谷歌命令数据集和音乐流派数据集上进行实验,结果表明,所提语音对抗样本生成方法的攻击成功率与主流方法相比,分别提高了3.47%、5.1%,平均信噪比提升了3.2、1.49 dB,该方法具有较好的攻击效果和语音质量。

基于弱监督学习的双分支结直肠病理图像腺体分割

现有弱监督分割方法难以获得结直肠病理图像的细粒度腺体特征,导致无法生成高质量伪标签的问题,影响腺体分割的效果。为了解决上述问题,提出一种基于弱监督学习的双分支结直肠病理图像腺体分割方法。首先,将patch级结直肠病理图像输入到第一个分支网络中,通过特征交互模块和亲和度注意力融合模块实现patch级图像的局部和全局特征的交互和融合,并获得细粒度腺体特征。然后,将图像级结直肠病理图像输入到第二个分支网络中,利用局部类激活注意力模块定位腺体位置,并获得粗粒度类激活图。最后,通过细粒度腺体特征和粗粒度类激活图,得到高质量伪标签,并在分割网络中经过跨尺度连接空间感知模块,实现腺体分割。实验结果表明,将所提方法在GlaS和CRAG两个结直肠病理图像数据集中进行实验,与其他分割方法相比取得较好的分割效果,验证所提方法的有效性。

基于度量学习的多分支舌象识别网络

为提升舌象识别效率与精准度,通过度量学习研究辅助医生识别舌象表征的方法。首先,收集舌诊图像111例,数据按照7:3的比例随机分为训练集和测试集。然后,设计一种基于度量学习的多分支舌象识别网络。深度学习网络被分为两个部分,前半部分为共享权重层,采用基于度量学习的舌象特征编码损失函数,以获得精准的特征;后半部分针对中医舌象的分类分为4个舌象识别辅助分支,降低舌象识别难度,提升准确率。此外,构建多标签残差映射,增加类间距,减小类内距,提升最终识别的准确度。本文方法在舌象数据集的测试集上进行测试时获得84.8%的识别精度,表明多分支网络架构可以很好地降低舌象识别难度,特别是特征类别较多的舌形和苔质。同时,舌象特征编码损失函数可以有效地提取舌象特征;舌象多标签残差映射可以减少各类别之间的干扰,从而提升识别准确度。

地图注记质量评价模型的研究

抽象出冲突、压盖、位置优先级和关联性4个基本因素,通过建立评分体系,构造地图注记的4因素形式化质量评价模型。同时,介绍该质量评价模型在地图自动注记系统MapLa bel中的应用情况和实验结果,说明此模型的合理性和有效性。