Regional cerebral blood flow imaging assessment of brain function reconstruction in elderly hemiplegia patients by body weight support treadmill training

The mechanism underlying body weight support treadmill training in elderly hemiplegic stroke patients is largely unknown. This study aimed to elucidate the changes of cortical blood flow in seven elderly patients with post-stroke hemiplegia before and after body weight support treadmill training by semi-quantitative analysis of regional cerebral blood flow assessed by single photon emission computed tomography. Body weight support treadmill training for 6 months was effective in improving cerebral blood flow and promoting the walking speed and balance recovery in elderly patients with post-stroke hemiplegia.

Intervention study of finger-movement exercises and finger weight-lift training for improvement of handgrip strength among the very elderly

Objectives:To examine the effects of finger-movement exercises and finger weight-lift training on handgrip strength and Activities of Daily Living Scale(ADLS)values.Methods:A total of 80 very elderly adults(aged
80 years)were assigned to either an intervention group(n?40)or a control group(n?40).Subjects in the intervention group performed finger-movement exercises and weight-lift training for a period of 3 months,while subjects in the control group received no intervention,and were unaware of the interventions received in the other group.Results:After completing 3 months of finger-movement exercises and weight-lift training,the average handgrip strength of the 40 participants in the intervention group had increased by 2.1 kg,whereas that in the control group decreased by 0.27 kg(P<0.05).After receiving intervention,the number of subjects in the intervention group with an ADLS score>22 points decreased by 7.5%(P<0.05,vs.pre-intervention).Conclusions:The combined use interventionwith finger-movement exercises and proper finger weight-lift training improved the handgrip strength andADLS values of very elderly individuals.These rehabilitation exercisesmay be used to help the elderlymaintain their self-care abilities.

Comparison of the Pattern of Weight Training on Muscle Strength and Endurance in Karate Men

Some studies performed weight training for improvement physical fitness in martial arts. The purpose of this study was to compare weight training with two loading patterns （double pyramid and reverse step） on muscular strength and muscular endurance in karate men. The subjects who have experienced six years in karate are participated voluntarily. These subjects have been divided randomly into three groups： double pyramid group, reverse step group and control group. Characteristics of subject： double pyramid group （age： 20.55 ± 2.62 years, weight： 73.45 ± 11.17 kg and height： 176.86 ±5.04 cm）, reverse step group （age： 20.50 ± 3.92 years, weight： 75.35 ± 11.24 kg and height： 178.90 ± 9.24 cm） and control group （age： 21.36 ±4.65 year ,weight： 75.77 ± 6.07 kg and height： 178.64± 4.67 cm）. All subjects performed weight training during eight weeks. Data were statistically analyzed by text dependent, analysis of variance and tukey test with a 0.05 significant level. The result shows that a program of weight training double-pyramid method significantly increases muscular strength and muscular endurance, while a period of weight training reverses step method only significantly increase muscle endurance but no change in muscular strength is observed. The coaches are recommended to use a double pyramid method to improve muscular strength and muscular endurance training with regard to training load for preparing their athletes.

Effects of Water Weight-Loss Walking Training on Lower Limb Motor Function and Gait in Stroke Patients

Background: Water weight-loss walking training is an emerging physical therapy technique, which provides new ideas for improving the motor function of stroke patients and improving the quality of life of patients. However, the rehabilitation effect of water weight-loss training in stroke patients is currently unclear. Objective: To analyze the effect of water weight loss walking training in stroke patients. Methods: A total of 180 stroke patients admitted to our hospital from January 2019 to December 2021 were selected and randomly divided into two groups. The control group received routine walking training, and the research group performed weight loss walking training in water on this basis. The lower limb motor function, muscle tone grade, daily living ability, gait and balance ability were compared between the two groups before and after treatment. Results: Compared with the control group, the FMA-LE score (Fugl-Meyer motor assessment of Lower Extremity), MBI score (Modified Barthel Index) and BBS score (berg balance scale) of the study group were higher after treatment, and the muscle tone was lower (P Conclusion: Water weight loss walking training can enhance patients’ muscle tension, correct patients’ abnormal gait, improve patients’ balance and walking ability, and contribute to patients’ motor function recovery and self-care ability improvement.

Research on Node Classification Based on Joint Weighted Node Vectors

Node of network has lots of information, such as topology, text and label information. Therefore, node classification is an open issue. Recently, one vector of node is directly connected at the end of another vector. However, this method actually obtains the performance by extending dimensions and considering that the text and structural information are one-to-one, which is obviously unreasonable. Regarding this issue, a method by weighting vectors is proposed in this paper. Three methods, negative logarithm, modulus and sigmoid function are used to weight-trained vectors, then recombine the weighted vectors and put them into the SVM classifier for evaluation output. By comparing three different weighting methods, the results showed that using negative logarithm weighting achieved better results than the other two using modulus and sigmoid function weighting, and was superior to directly concatenating vectors in the same dimension.

一种基于Tri-training的数据流集成分类算法

数据流分类是数据挖掘领域的重要研究任务之一,已有的数据流分类算法大多是在有标记数据集上进行训练,而实际应用领域数据流中有标记的数据数量极少。为解决这一问题,可通过人工标注的方式获取标记数据,但人工标注昂贵且耗时。考虑到未标记数据的数量极大且隐含大量信息,因此在保证精度的前提下,为利用这些未标记数据的信息,本文提出了一种基于Tri-training的数据流集成分类算法。该算法采用滑动窗口机制将数据流分块,在前k块含有未标记数据和标记数据的数据集上使用Tri-training训练基分类器,通过迭代的加权投票方式不断更新分类器直到所有未标记数据都被打上标记,并利用k个Tri-training集成模型对第k+1块数据进行预测,丢弃分类错误率高的分类器并在当前数据块上重建新分类器从而更新当前模型。在10个UCI数据集上的实验结果表明:与经典算法相比,本文提出的算法在含80%未标记数据的数据流上的分类精度有显著提高。

改进M-training算法的高光谱图像分类

为了解决高光谱数据有标签样本数量有限的分类问题,提出将M-training算法应用于高光谱图像分类。采用两个SVM、一个K近邻(KNN)以及一个随机森林(RF)进行分类器组合,对传统M-training算法进行改进,增强分类器的多样性和差异性。为了充分考虑大量无标签样本的影响,采用有标签样本与无标签样本错误率加权作为有标签样本集更新的限制条件,从而有效地扩大了有标签样本集。实验结果表明:改进算法和传统的M-training算法相比较,在总体分类精度与Kappa系数上分别提高1. 85%～12. 10%与0. 021 5～0. 141 3,从而验证了该算法的有效性。

一种改进的图像边缘检测方法——Weighted CHNN

文中提出了一种改进的CHNN方法,称为Weighted CHNN(加权的CHNN,简称WCHNN)方法.该方法在CHNN神经网络元的n个连接上施加权值,可以通过各种局部搜索、优化算法,使用指定的样本输入、样本输出等方法来训练该WCHNN网络从而确定各权值,使得WCHNN在保留了CHNN的优点的同时,还可以根据不同的样本输入输出图像来调节边缘检测的灵敏度,从而提高检测结果质量并避免检测结果中出现边缘过宽的情况.实验结果表明,训练后的WCHNN网络,比起CHNN有着更低的边缘检测错误率,并可检出原来CHNN方法漏检的边缘.

一种改进的CHNN图像边缘检测方法—Weighted CHNN

针对文献[1]中提出的CHNN图像边缘检测算法缺乏足够的参数来调节边缘检测的灵敏度以及检测结果图像边缘过宽的缺陷,提出一种改进的CHNN方法,称之为Weighted CHNN(加权的CHNN,简称WCHNN)方法。该方法在CHNN神经网络元的n个连接上施加权值,可以通过各种局部搜索、优化算法,使用指定的样本输入、样本输出等方法来训练该WCHNN网络从而确定各权值,使得WCHNN在保留了CHNN的优点的同时,还可以根据不同的样本输入输出图像来调节边缘检测的灵敏度,从而提高检测结果质量并避免检测结果中出现边缘过宽的情况。实验结果表明,训练后的WCHNN网络,比起CHNN有着更低的边缘检测错误率,并可检出原来CHNN方法漏检的边缘。

The relationship of high-intensity cross-training with arterial stiffness

Background'.Central arterial stiffness is a cardiovascular risk factor that can be readily affected through engagement in physical exercise training,with resistance and aerobic exercise having disparate affects.Despite the growing popularity of high-intensity cross-training(HICT),little is currently known about the effects of this mixed modality exercise stimulus on arterial stiffness.Therefore,the purpose of this study was to characterize the arterial stiffness of habitual HICT participants vs.aerobically active and sedentary controls using a cross-sectional design.Methods'.A total of 30 participants were recruited:10 middle-aged long-term participants of HICT(CrossFit)and 20 age,sex,and height matched controls(10 recreationally active,10 sedentary).Central and peripheral pulse wave velocities were measured for the carotid-femoral and femoral-dorsalis pedis arterial segments.Aerobic fitness(maximal oxygen uptake,VChmax)was measured and typical exercise participation rates were self-reported for each group.Results'.HICT participants manifested central pulse wave velocity(PWV)(5.3±1.0 m/s,mean±SD)and VChmax(43±6 mL/kg/min)values nearly identical to active controls.Both active groups had significantly better values than sedentary controls(7.1±1.0 m/s,p≤0.001;and 32±7 mL/kg/min,p=0.01).No differences were observed in peripheral PWV between groups.Conclusion'.Habitual participation in HICT exercise was not associated with increased central nor peripheral arterial stiffness.Long-term HICT participants presented with similar fitness and arterial stiffness as compared with participants who practiced traditional aerobic exercise.Compared to sedentary living,HICT may offer musculoskeletal and cardiovascular health benefits without negatively impacting arterial stiffness.

作业疗法联合减重平板步行训练对脑卒中偏瘫患者平衡与功能性步行能力的影响

目的:探讨作业疗法联合减重平板步行训练对脑卒中偏瘫患者平衡与功能性步行能力的影响。方法:选取2019年7月—2021年6月河南科技大学第一附属医院新区医院康复医学科收治的90例脑卒中偏瘫患者作为研究对象,采用随机数表法分为试验组(常规康复护理)与对照组(作业疗法联合联合减重平板步行训练),每组各45例。比较两组患者干预前及干预3个月的平衡功能与功能性步行能力。结果:干预3个月,试验组患者Berg平衡量表(BBS)、功能性步行量表(FAC)评分显著高于对照组,差异有统计学意义(t=23.279、5.555,P<0.05);试验组患者脑卒中专用生活质量量表(SS-QOL)各项评分显著高于对照组,差异有统计学意义(t=8.424、8.637、26.306、5.162、7.579、15.302、8.918、8.080、20.745、10.862、1.397、6.806,P<0.05)。结论:作业疗法联合减重平板步行训练可有效提高脑卒中偏瘫患者平衡功能及功能性步行能力,有利于改善患者的生活质量。

基于博弈论组合赋权-云模型的消防员训练损伤风险评估

为降低消防员训练损伤发生率,从个人因素、环境因素和组织管理因素3个方面,提出包含19个指标的消防员训练损伤风险评估指标体系;采用Bonferroni有序加权平均算子(BON-OWA算子)修正层次分析法(AHP)的误差,获得各指标的主观权重;引用基于指标相关性的权重确定法(CRITIC)确定各指标的客观权重;基于博弈论组合赋权法确定各指标的组合权重;引入云模型理论,建立消防员训练损伤风险评估模型,并以河南省某消防救援大队消防员的训练为例,验证该模型的科学性。结果表明:训练装备的战技性能状况指标对消防员的训练损伤风险等级影响最大;该风险评估模型通过数据处理,可以确定消防员训练损伤风险等级;案例中,消防员在5种不同情景下的训练损伤风险等级均为一般风险,这与模糊综合评价法所得结果一致。

基于Vague集犹豫量化的机舱团队协作式评估方法

针对现有机舱团队协作评估方法不统一且主观随意性太强的缺陷,提出一种基于Vague集犹豫量化的新型评估方法。从机舱团队协作的角度对影响评估的各因素进行提取,采用集合相似系数对重要协作指标量化,并采用C-OWA算子赋权法和变异系数法确定主客观权重,利用最小鉴别信息原理求得组合权重,结合Vague集的犹豫量化特点,通过机舱团队实际操作数据与Vague集评估标准得到各指标的定性评估Vague集,并利用集合相似度分析实现机舱团队协作的综合评估。实例计算结果表明:相对于传统的评估方法,在个体完成度的基础上加入具有犹豫团队协作因素的方法,能对评估结果的犹豫度进行较好地刻画,也能很好地实现定量与定性之间的转化,是一种有效且优越的机舱团队协作式评估方法。

基于DEMATEL和SE-DEA的民航监察员培训效能评价

为优化培训模式,提高监察员的安全监管能力,降低民航各种不安全事件发生的概率,以“岗位胜任能力导向”为核心的培训模式逐步形成。首先基于“一个素质,三个能力”构建初始岗位胜任力模型,然后结合SPSS和AMOS进行指标筛选和模型验证,利用DEMATEL法探究各指标之间的影响关系并改进权重,最后运用SE-DEA法检验目前的培训课程设置是否合理并进行评价。结果表明:构建的岗位胜任力共15项指标,具有良好的效度和信度;政治素质和管理能力对法治能力和业务能力的影响更大,对胜任力模型影响最大的是业务能力;经过培训,75%的监察员的岗位能力都得到有效提升,达到强化基层安全能力建设的预期,同时说明目前的培训课程设置仍需进一步改进。

基于加权融合的常导高速磁浮列车UKF定位算法

为提高高速磁浮列车测速定位的精确性,本文针对基于长定子齿槽检测的常导高速磁浮列车测速定位方法在列车运行过程中可能因测速定位信号缺失、干扰、测速定位安装误差等原因引起的定位不准问题,提出一种基于加权融合无迹卡尔曼滤波(UKF)的常导高速磁浮列车测速定位算法.介绍了高速磁浮列车基于长定子齿槽的测速定位方法,并对多路冗余速度位置信息进行预处理和自适应加权融合处理;给出基于加权融合UKF的常导高速磁浮列车测速定位算法模型;基于磁浮列车测速定位在环测试试验台试验,对改进后的无迹卡尔曼滤波磁浮定位算法与原定位算法进行了对比分析.分析结果表明:磁浮列车平均速度误差减小了32.6%,速度极差降低了49.3%,有效消除了信号采集噪声,提高了磁浮列车测速定位精度.

一种测控装备训练成效评估指标体系构设

文章以雷测装备跟踪无人机训练为主要手段,为有效评估操作手能力水平,从可操作性的角度出发,选取反应时长、总跟踪有效时长、测角精度三个指标要素作为跟踪训练成效评估指标,并用层次分析法求出各指标的权重。该评估指标体系充分遵循了客观性、简明性、独立性和可测性等原则,可消除人为因素的影响,为雷测装备捕获跟踪训练成效评估模型的建立奠定了基础,具有一定的实用价值。

果树枝条识别与修剪点坐标确定方法

为了实现休眠期果树的修剪识别问题,研究了基于语义分割的网络模型识别修剪枝条与确定修剪点坐标的方法。通过双目相机搭建了视觉系统获取果树的数据集,采用分别融入预训练权重与CBAM(注意力机制)的VGG16和RestNet-50作为U-Net主干特征提取网络的2种深度学习模型分割修剪枝条,同时获取其效果并进行对比。在获得的分割图像基础上采用骨架提取和修剪点聚类2种方法进行修剪点坐标的确定。结果表明,基于VGG16特征提取网络的U-Net模型识别结果较好,该模型在测试集的平均交并比(MIOU)、平均像素准确率(MPA)和训练时F分数分别为84.80%、91.83%和92.679%。分割出人工模拟果树的模型图像,采用修剪点聚类的方法,可以较快、实时地确定修剪点的二维坐标,为实现修剪作业奠定基础。

适应人体重心起伏的悬吊减重康复系统设计

减重系统对下肢疾病患者的行走康复训练有重要影响。现有的下肢康复外骨骼减重装置大多只考虑如何减去患者体重的百分比,而忽略了患者身体重心的起伏。由于外骨骼的骨盆支架在竖直方向运动轨迹固定,患者步态的微小变化就会导致重心高度与骨盆支架运动轨迹不符,这一差异会作用到患者的骨盆位置,影响下肢关节的活动,并产生额外的风险。针对这一问题,提出通过采集足底压力来预测重心位置变化,并利用获得的重心轨迹解算出应施加的减重力方法,为患者训练提供安全有效的减重系统。所提方法的可行性已获仿真证实,并在研发的外骨骼减重系统上获得实际验证,模糊控制器跟踪重心轨迹的误差相比PID控制减少了21.2%,稳态误差维持在±1 mm范围内,髋膝关节的活动范围相比常规减重分别增加了14.36%和13.77%。

基于权重复用的训练加速算法

深度学习已广泛应用在科研教学、工业生产等各领域,但因其数据量庞大和模型结构复杂,在模型训练阶段要依赖大量的计算资源。为了能在实验教学环节提升资源利用效率,让学生更加熟练地掌握数据搜集和模型参数调整优化的能力,提出了一种基于权重复用的训练加速方法,分别对VGG和ResNet网络结构的深度和宽度进行伸缩拓展,允许模型复用结构相似但不需要完全一致的网络权重。实验结果表明,在CIFAR10数据集上测试,采用权重复用方法进行初始化的训练更快收敛,而且在训练结束时与随机化训练的准确率相近,实现拓展后的网络加速训练,是一种更加灵活的知识迁移方法,有助于培养学生对复杂模型的理解与优化能力。

基于字句动态特征和自注意力的情感分析方法

针对评论情感分析任务中文本长度失衡引起的特征稀疏、特征缺失和提取信息不全等问题,提出了一种基于字句动态特征和自注意力的情感分析方法。首先基于预训练模型对评论进行动态特征编码,使用句向量对不足固定长度的部分进行补全,并表征超出的截断部分,以缓解批训练下文本尺寸失衡引起的特征稀疏和特征缺失问题。然后使用基于自注意力机制的特征重组方法动态整合字句融合特征,并优化权重参数以降低计算和训练时间复杂度。最后在开源数据集上分别进行了消融实验和对比实验。测试结果表明,本文方法在准确率上有较优的改进效果。