SGT-Net: A Transformer-Based Stratified Graph Convolutional Network for 3D Point Cloud Semantic Segmentation

In recent years,semantic segmentation on 3D point cloud data has attracted much attention.Unlike 2D images where pixels distribute regularly in the image domain,3D point clouds in non-Euclidean space are irregular and inherently sparse.Therefore,it is very difficult to extract long-range contexts and effectively aggregate local features for semantic segmentation in 3D point cloud space.Most current methods either focus on local feature aggregation or long-range context dependency,but fail to directly establish a global-local feature extractor to complete the point cloud semantic segmentation tasks.In this paper,we propose a Transformer-based stratified graph convolutional network(SGT-Net),which enlarges the effective receptive field and builds direct long-range dependency.Specifically,we first propose a novel dense-sparse sampling strategy that provides dense local vertices and sparse long-distance vertices for subsequent graph convolutional network(GCN).Secondly,we propose a multi-key self-attention mechanism based on the Transformer to further weight augmentation for crucial neighboring relationships and enlarge the effective receptive field.In addition,to further improve the efficiency of the network,we propose a similarity measurement module to determine whether the neighborhood near the center point is effective.We demonstrate the validity and superiority of our method on the S3DIS and ShapeNet datasets.Through ablation experiments and segmentation visualization,we verify that the SGT model can improve the performance of the point cloud semantic segmentation.

SEGMENTIZED OPTIMIZATION STRATEGY FOR PREDICTIVE CONTROL

To improve the computational efficieney of optimization based control methods, a new kind of Segmentized Optimization Strategy is presented,aiming at achieving more economical computation as well as comparatively satisfactory performance. Its profitability is examined. And the effectiveaess is shown in the simulation.

Invasive versus conservative strategy in consecutive patients aged 80 years or older with non-ST-segment elevation myocardial infarction: a retrospective study in China

Objective To investigate whether the very elderly patients with non-ST-segment elevation myocardial infarction (NSTEMI) will benefit from an invasive strategy versus a conservative strategy. Methods 190 consecutive patients aged 80 years or older with NSTEMI were included in the retrospective study from September 2014 to August 2017, of which 69 patients received conservative strategy and 121 patients received invasive strategy. The primary outcome was death. Multivariate Cox regression models were used to assess the statistical association between strategies and mortality. The survival probability was further analyzed. Results The primary outcome occurred in 17.4% patients in the invasive group and in 42.0% patients in the conservative group (P = 0.0002). The readmission rate in the invasive group (14.9%) was higher than that in the conservative group (7.2%). Creatinine level (OR = 1.01, 95% CI: 0.10–1.03, P = 0.05) and use of diuretic (OR = 3.65, 95% CI: 1.56–8.53, P = 0.003) were independent influential factors for invasive strategy. HRs for multivariate Cox regression models were 3.45 (95% CI: 1.77–6.75, P = 0.0003), 3.02 (95% CI: 1.52–6.01, P = 0.0017), 2.93 (95% CI: 1. 46–5.86, P = 0.0024) and 2.47 (95% CI: 1.20–5.07, P = 0.0137). Compared with the patients received invasive strategy, the conservative group had remarkably reduced survival probability with time since treatment (P < 0.001). Conclusions An invasive strategy is superior to a conservative strategy in reducing mortality of patients aged 80 years or older with NSTEMI. Our results suggest that an invasive strategy is more suitable for the very elderly patients with NSTEMI in China.

An Improved Unsupervised Approach to Word Segmentation

ESA is an unsupervised approach to word segmentation previously proposed by Wang, which is an iterative process consisting of three phases: Evaluation, Selection and Adjustment. In this article, we propose Ex ESA, the extension of ESA. In Ex ESA, the original approach is extended to a 2-pass process and the ratio of different word lengths is introduced as the third type of information combined with cohesion and separation. A maximum strategy is adopted to determine the best segmentation of a character sequence in the phrase of Selection. Besides, in Adjustment, Ex ESA re-evaluates separation information and individual information to overcome the overestimation frequencies. Additionally, a smoothing algorithm is applied to alleviate sparseness. The experiment results show that Ex ESA can further improve the performance and is time-saving by properly utilizing more information from un-annotated corpora. Moreover, the parameters of Ex ESA can be predicted by a set of empirical formulae or combined with the minimum description length principle.

Enhanced Chimp Optimization Algorithm Using Attack Defense Strategy and Golden Update Mechanism for Robust COVID-19 Medical Image Segmentation

Medical image segmentation is a powerful and evolving technology in medical diagnosis.In fact,it has been identified as a very effective tool to support and accompany doctors in their fight against the spread of the coronavirus(COVID-19).Various techniques have been utilized for COVID-19 image segmentation,including Multilevel Thresholding(MLT)-based meta-heuristics,which are considered crucial in addressing this issue.However,despite their importance,meta-heuristics have significant limitations.Specifically,the imbalance between exploration and exploitation,as well as premature convergence,can cause the optimization process to become stuck in local optima,resulting in unsatisfactory segmentation results.In this paper,an enhanced War Strategy Chimp Optimization Algorithm(WSChOA)is proposed to address MLT problems.Two strategies are incorporated into the traditional Chimp Optimization Algorithm.Golden update mechanism that provides diversity in the population.Additionally,the attack and defense strategies are incorporated to improve the search space leading to avoiding local optima.The experimental results were conducted by comparing WSChoA with recent and well-known algorithms using various evaluation metrics such as Feature Similarity Index(FSIM),Structural Similarity Index(SSIM),Peak signal-to-Noise Ratio(PSNR),Standard deviation(STD),Freidman Test(FT),and Wilcoxon Sign Rank Test(WSRT).The results obtained by WSChoA surpassed those of other optimization techniques in terms of robustness and accuracy,indicating that it is a powerful tool for image segmentation.

A critique of Porter＇s cost leadership and differentiation strategies

Porter identifies high market share with cost leadership, citing GM as a successful practitioner of this strategy. However, GM became a market share leader in the American automobile industry due to a strategy of market segmentation, differentiation and a broad scope shaped during the 1920s. Porter argues that cost leadership and differentiation offer an equally viable path to competitive success. Nevertheless, a differentiation strategy based on superior quality compared to competition is more profitable than cost leadership strategy. It can lead a business to become a market share leader, and consequently even a low-cost leader. Research indicates that differentiation and cost leadership can co-exist. However, Porter insists that each generic strategy requires a different culture and a totally different philosophy. The problem is that Porter＇s generic strategies are too broad. It is not his logic that is flawed, but his basic premise that prescribes cost leadership strategy as the only route to market share leadership, and presents a narrow view of differentiation with a unique product--sold at a premium price--on the one hand, and a ＂standard, or no-frills＂ product on the other. Mintzburg （1988） says Porter＇s cost leadership strategy should be called ＂price differentiation＂： a strategy that is based on a lower price than that of the competition. He suggests that business strategy has two dimensions： differentiation and scope. Thus, setting scope aside, competitive strategy has only one component： differentiation. So, the key question is not whether to differentiate, but how？ First, make customer-perceived quality as the foundation of competitive strategy because it is far more critical to long-term success than any other factor. Second, serve the middle class by competing in the mid-price segment, offering better quality than the competition at a somewhat higher price. It is this path that can lead to market share leadership--a strategy that can be both profitable--and sustainable.

Perspective of BiH Tourism as a Segment of World Tourism

Tourism is one of the most important drivers of economic development in the world,and probably the only activity that connects peoples and countries without any prejudice.This is one of the main reasons why tourism is developing at a faster pace and is becoming a more modern activity for man.Proper positioning in a dynamic and competitive tourism market is conditioned by the monitoring of current trends that identify demand trends,new market opportunities,areas of potential investment,and infrastructure needs.Tourism as a business promises much for the whole economy.Man is increasingly investing his efforts to improve both supply and demand in the sector.The tourism development strategy deals with the future as a long-term tourism development plan has already been drawn up at least several years in advance.The offer is more modern and anyone who wants to see,learn,recreate,or just take a good break comes across a wide range of offers.Tourism used to be only for the selected class of society,and today and in the future it will provide much to the average person.Today’s world trend allows for what was once unimaginable,so today travel agencies have to offer New Year’s Eve in a spacecraft,and who knows what else the future of tourism has in store for us?The purpose of our work is that tourism must become one of the main drivers of development in the areas of BiH in which it has so far been marginalized and use its full potential.The entire territory of BiH has an extremely diverse and preserved natural and cultural tourism potential,which is a valuable legacy that must be preserved and protected in order to contribute to tourism development in the long run.The aim of this paper is to create an image for the development of a quality,modern,and innovative tourism offer,to get acquainted with the novelties in the world tourism offer,and thus to bring BiH into the leading tourist countries.The next chapter deals with the vision of tourism in the future,the views on megatrends,and their reflection on the future development of tourism.The paper also deals with the recent changes in the supply and demand market,describes the formation and detection of segmentation of the tourism market.In this paper we will try to bring you closer to what destination management is and its purpose as well as tourism trends of the future.Selective forms of tourism,cultural tourism,long-distance tourism,alternative tourism,space tourism will be addressed,up to the modern accommodation offer.

退役动力电池双层均衡方法研究

针对如何提高退役电池组均衡速度,提出了一种双层均衡电路,底层电路采用基于电感的Buck-Boost均衡电路结构,顶层电路采用可重构电路结构。根据开路电压(OCV)和电池荷电状态(SOC)曲线关系采用分段均衡控制策略。设计双层均衡电路的仿真实验,仿真实验结果表明:在电池参数相同情况下,与传统Buck-Boost均衡电路相比,双层电路均衡时间减少25%,有效提高了电池组均衡速度。在电池状态和均衡电路结构相同的情况下,与单一变量的均衡策略相比,SOC-电压分段均衡策略时间减少11%,验证了该均衡方案可行性。

特长水下公路盾构隧道结构防火保护策略及关键参数研究

特长水下公路盾构隧道普遍设置独立排烟道进行顶部集中排烟。现有隧道工程防火技术“护顶不护墙”,顶部设有排烟道时,顶部防火保护施加在烟道板上,而烟道层拱顶管片及车道层两侧管片不施加防火保护,存在安全隐患。根据隧道结构实际可能受到的火灾高温作用,结合管片、烟道板结构相对重要性,提出了新的车道层侧墙管片、烟道层拱顶管片、烟道板结构防火保护策略,即车道层所有结构表面采用RABT火灾升温曲线,烟道层所有结构表面采用基于CFD模拟的火灾升温曲线,耐火极限均不小于2.0h,管片耐火极限由管片混凝土表面及距表面25mm处温度判定,烟道板耐火极限由烟道板的极限弯曲变形量和极限弯曲变形速率判定。基于所提出的防火保护策略,开展数值模拟及耐火试验研究,确定了防火保护技术参数。结果表明:车道层侧墙管片防火保护层等效热阻应不小于0.10℃·m~2/W;烟道层拱顶管片防火保护层等效热阻在排烟口处不小于0.10℃·m~2/W,且随距排烟口距离的增大逐渐减小,距离排烟口16m处减小为零;隧道侧墙常用装饰材料搪瓷钢板在火灾过程中易发生变形而脱落;无防火保护烟道板耐火极限大于2.0h,掺加PP纤维、钢纤维可提升烟道板的耐火性能。

基于SWOT的润滑油集团客户营销策略分析

润滑油企业要针对集团客户的行业典型特点和潜在需求,定制客户市场营销策略。基于客户价值和满意度特征等,可以将集团客户市场进一步细分为四类。SWOT分析将与润滑油市场营销中密切相关的各种主要内部优势、劣势和外部的机会及威胁等,通过调查列举出来,并依照矩阵形式排列。6Ps市场营销组合是六方面因素的组合。针对润滑油市场营销的特点,将6Ps营销组合理论和SWOT分析结合起来,制定SO-A、WO-A、ST-A、WT-A等十六类策略,这些策略的灵活运用为润滑油企业适应新时代转型发展提供了理论依据。

基于多标签语义分割的硬笔字笔画提取

汉字作为中华文化的载体,因其复杂的结构区别于其他文字.笔画作为汉字的基本单元,在硬笔字评价中起到至关重要的作用.正确提取笔画,是硬笔字评价的首要步骤.现有的笔画提取方法多数是基于规则的,由于汉字的复杂性,这些规则通常无法顾及所有特征,且在评价时无法根据笔顺等信息与模板字笔画匹配.为了解决这些问题,该文将笔画提取转化为多标签语义分割问题,提出了多标签语义分割模型(M-TransUNet),利用深度卷积模型以汉字为单位任务进行训练,保留了笔画原有结构,避免了笔画段组合的二义性,同时得到了硬笔字的笔顺,有利于笔画评价等下游任务.由于硬笔字图像只分为前景和背景,没有额外颜色信息,所以更容易产生FP(false positive)分割噪声.为解决此问题,本文还提出了一种针对笔画分割结果的局部平滑策略(local smooth strategy on stroke,LSSS),淡化噪声的影响.最后,本文对M-TransUNet的分割性能以及效率进行了实验,证明了本文算法在很小性能损失的情况下,极大地提升了效率.同时对LSSS算法进行了实验,证明其在FP噪声消除的有效性.

基于门控机制与重放策略的持续语义分割方法

基于深度神经网络的语义分割模型在增量更新知识时由于新旧任务参数之间的干扰加之背景漂移现象,会加剧灾难性遗忘。此外,数据常常由于隐私、安全等因素无法被存储导致模型失效。为此,该文提出基于门控机制与重放策略的持续语义分割方法。首先,在不存储旧数据的情况下,通过生成对抗网络生成及网页抓取作为数据来源,使用标签评估模块解决无监督问题、背景自绘模块解决背景漂移问题;接着,使用重放策略缓解灾难性遗忘;最后,将门控变量作为一种正则化手段增加模型稀疏性,研究了门控变量与持续学习重放策略结合的特殊情况。在Pascal VOC2012数据集上的评估结果表明,在复杂场景10-2,生成对抗网络(GAN)、Web的设置中,该文在全部增量步骤结束后的旧任务性能比基线分别提升了3.8%,3.7%,在场景10-1中,相比于基线分别提升了2.7%,1.3%。

基于Sentinel-2影像的喀斯特地区土地利用信息提取

针对喀斯特地区分类尺度难以确定,特征数量维数过高,分类精度较低的问题,文章提出了通过联合评价确定最优分割尺度、ReliefF算法对先验特征数据集进行优选,使用分层掩膜的策略,利用随机森林算法完成分类的方法。并以贵阳西南喀斯特地区为研究区,首先使用同质性与Moran's I联合评价的方法确定最优分割尺度为80,通过ReliefF算法优选出重要性较高的15个特征;在此基础上,通过对比试验验证了随机森林算法的优越性;以2020年Sentinel-2影像为实验数据设计3种面向对象分类方案。结果表明,经最优尺度计算、特征优选和分层掩膜的分类方法结果精度最高,分类总体精度、Kappa系数、AD、QD分别达到0.886、0.849、0.092、0.022。最后将该方法应用于2023年影像,分类总体精度、Kappa系数、AD、QD分别达到0.868、0.825、0.106、0.026。证明了该方法在喀斯特地区土地利用信息提取方面的优越性和适用性。

结直肠癌免疫组化图像分级诊断方法

人体组织病理学检查主要应用于各类肿瘤诊断和治疗,免疫组织化学技术在结直肠癌早期筛查中具有重要的临床意义。为准确判断结直肠癌抑癌基因p53表达程度,文中提出一种基于逐块释放微调策略迁移学习的分级诊断方法,通过图像预处理、有监督模型预训练以及逐块释放微调等步骤将细胞核分割模型的参数迁移至诊断框架中。生成的细胞核分割掩膜进行主成分分析(Principal Component Analysis,PCA)降维和支持向量机(Support Vector Machine,SVM)多元分类最终得到图像诊断结果。该方法在结直肠癌p53蛋白免疫组化(Immunohistochemistry,IHC)图像数据集上进行了验证,模型的Dice值可达到0.8876,分级准确率达到90.28%。结果表明,所提方法能够对结直肠癌免疫组化图像有效分级,为医生阅片提供可靠的辅助信息。

Numerical Study of Water Transport and Operating Strategy for Fuel Cells with Segmented Water Management Flow Fields

The effective cathode flow field design can realize the internal water balance and higher current density output of proton exchange membrane fuel cells(PEMFC).Therefore,a segmented water management flow field is proposed in this study,i.e.a half separated-half coupled cathode(HSHC)flow field which has two inlets but just one outlet.A 3D numerical PEMFC model is applied to study the effect of the HSHC flow field on PEMFC performance and its operating strategy in terms of operating conditions.The study results are shown as follows:Compared with the two conventional cathode flow fields,the HSHC flow field improves the water balance along the channel and increases the current density by 17.1%at a cathode stoichiometry of 3.25.It is because the HSHC flow field can overcome the water loss at channels upstream and the water accumulation at channels downstream.The draw water phenomenon(DWP)in the HSHC flow field is observed,which is mainly affected by the water vapor pressure of channel.Based on the DWP,cooling channel inlet flow rates can be used to adjust water balance,but severe water loss should be avoided.In addition,the inlet temperature control in HSHC flow field should be that cell temperature>cathode channel inlet temperature>cooling channel inlet temperature>ambient temperatures for better water balance.

多策略融合改进北方苍鹰算法的森林冠层图像分割

针对森林冠层图像复杂,分割精度较差等问题,提出一种多策略融合改进北方苍鹰优化算法(Improved Northern Goshawk Optimization,INGO)的冠层图像分割方法。首先,在北方苍鹰初始化引入随机反向学习策略,以增加种群多样性,提高搜索效率;在北方苍鹰探索阶段,添加自适应权重因子提高算法搜索能力,加快算法收敛速度;在北方苍鹰开发阶段,添加非线性收敛因子平衡全局搜索和局部开发能力。其次,采用多阈值Kapur熵作为适应度函数,通过选取8个基准函数对改进的算法进行测试,测试结果表明,改进的算法可以有效提高精度和搜索速度。最后,将所改进的算法对森林冠层图像进行阈值分割试验,并在适应度值、森林图像分割时的峰值信噪比(PSNR)、结构相似性(SSIM)与特征相似度(FSIM)上进行对比分析。试验结果表明,改进的算法可以获得更精确的森林冠层分割阈值和更高的分割精度。

基于改进二进制蛇优化算法的配电网故障定位

分布式电源(distributed generation,DG)大规模接入给配电系统带来更多不确定性、随机性,系统运行方式更复杂,传统故障定位方法难以适应新型电力系统构建。提出了一种基于改进二进制蛇优化算法(improved binary snake optimization,IBSO)的新型故障区段定位方法。利用SPM混沌映射生成高质量的随机数序列,以提高算法种群中个体的随机性,并引入了遗传算法的动态变异策略,根据不同的搜索状态和进化阶段来调整变异率和变异方式,提高算法的灵活性和准确性。通过仿真证明,该方法适用于在含有分布式电源的配电网中定位单一和多重故障区段,相比蛇优化算法、传统二进制粒子群算法以及遗传算法在收敛性、快速性和准确性方面更优。

基于缓步升温的高精度恒温箱分段检测控制研究

温度控制是许多科学实验和工业生产的关键参数,精确温度控制对于实验结果的准确性和工业产品的质量至关重要。文章旨在研究一种基于缓步升温的温度控制方案,实现对温度的高精确控制。该方案采用缓步升温策略,通过分段检测技术实时监测温度变化,并通过嵌入式系统进行数据处理和判断。通过分段检测技术,该设备能够实时监测温度变化,并在异常情况下及时报警。实验结果表明,该方法能够有效提高恒温箱的温度控制精度,降低能耗,并延长设备使用寿命。

基于锂电池电压片段的退役电池快速筛选重组

针对退役电池梯次利用筛选精度差、效率低的问题,提出一种基于神经网络-支持向量机-线性回归(multilayer perceptron-support vector regression-logistic regression,MLP-SVR-LR)混合模型的退役电池筛选方法,基于新的电压片段特征提取方法从恒流充电电压片段中提取与容量相关的特征,结合MLP-SVR-LR混合模型,实现高精度和高效率的退役电池筛选;同时,提出一种基于粒子群改进K-means算法和以轮廓系数(silhouette coefficient,SC)为标准的改进等数策略对筛选后的电池以容量和内阻为标准进行均分重组,实现退役电池直接成组利用。搭建磷酸铁锂电池仿真和实验平台,对所提电池筛选、重组策略进行实验和仿真分析。结果表明,所提方法较同类方法筛选精度提升2%以上,从退役电池筛选精度、筛选效率和重组后电池一致性的角度验证了所提方法的正确性和有效性。

基于角度信息的1比特RIS波束赋形

智能超表面(Reconfigurable Intelligent Surface,RIS)技术是第六代移动通信网络(6th generation mobile networks,6G)的关键支撑技术之一,可显著增强信号强度和质量,减轻通信干扰,提升通信性能;相较于传统有源天线阵列,RIS硬件成本低,被动工作方式带来低功耗的特性。波束赋形技术是RIS辅助通信系统的关键技术之一,过去的研究主要集中在连续相位波束赋形上,解决典型的通信波束赋形目标,包括能量效率最大化、信噪比最大化以及传输速率最大化等。这些研究取得了显著的成果,为通信系统的性能提升做出了贡献。然而,在实际应用中,由于成本和硬件方面的限制,离散波束赋形设计更受青睐。现有的离散波束赋形方案难以保证最优性,且依赖于高开销的RIS级联信道估计。在这一背景下,本文提出了两种1比特RIS波束赋形方法,包括基于分割的最优波束赋形方法和基于贪婪策略的低复杂度波束赋形方法。这两种方法分别通过巧妙的分割和迭代优化,实现了只需角度信息即可生成指定方向高增益波束的目标。具体而言,基于分割的波束赋形方法通过复平面分割的方式,巧妙地设计各RIS单元的相位,可获得最优解;基于贪婪策略的波束赋形方法,通过对各RIS单元相位迭代优化的方式,寻找1比特相位配置,能够取得接近最优的性能且具有更低的复杂度。此外,随着RIS相位误差的增加,分割法相对于贪婪法略显优势。