表面肌电信号采集与降噪处理

目的：本文以设计的表面~g（sEMG）信号采集系统为基础,探讨sEMG信号中的降噪处理问题。方法：结合sEMG信号的噪声影响情况,首先利用带通滤波器消除肌电信号频带外噪声,再通过频谱插值法来抑制工频干扰分量,最后使用小波分析方法来削弱肌电信号频带内噪声。结果：通过对检测sEMG信号的降噪处理,信号噪声得到明显抑制。结论：所设计采集系统能够获得满意的sEMG信号检测效果,所采用降噪方法能够有效提高sEMG信号的质量。     

表面肌电信号的肌肉疲劳判别研究进展

肌肉在周期的收缩或静态的拉伸过程中,会渐渐进入疲劳状态,肌肉疲劳特性的研究在康复医学、运动医学领域具有重要作用。表面肌电信号是从肌肉表面通过电极记录下来的反映神经肌肉系统活动的一维时间序列非平稳生物电信号,是评价局部肌肉疲劳的有效工具。本研究从时域和频域、时频域线性方法下的测量指标和非线性方法下的指标来综述表面肌电信号的疲劳研究进展,同时比较各种方法的优缺点,并对使用表面肌电信号来判别疲劳研究做了进一步的展望。     

肌肉动态收缩期间表面肌电信号的时频分析

介绍了用于肌肉动态收缩期间非平稳表面肌电信号的时频分析方法。用短时傅里叶变换、Wigner-Ville分布及Choi-Williams分布计算了表面肌电信号的时频分布,用于信号频率内容随时间演化的可视化观察。通过计算瞬时频谱参数,对肌肉疲劳的电表现进行量化描述。分析了反复性的膝关节弯曲和伸展运动期间从股外侧肌所记录的表面肌电信号。发现和在静态收缩过程中观察到的平均频率线性下降不同,在动态收缩期间瞬时平均频率的变化过程是非线性的并且更为复杂,且与运动的生物力学条件有关。研究表明将时频分析技术应用于动态收缩期间的表面肌电信号可以增加用传统的频谱分析技术不能得到的信息。     

局部肌肉疲劳的表面肌电信号复杂度和熵变化

目的 在于探讨静态和动态疲劳性运动过程中肱二头肌和腰部脊竖肌表面肌电(surface electromyography,sEMG)信号的Lempel-Ziv复杂度和Kolmogorov熵的变化规律。18名男性大学生志愿者被随机分为肱二头肌和腰部脊竖肌运动负荷组,分别完成静态和动态疲劳运动负荷试验。运动负荷期间连续记录sEMG信号,在对运动负荷时间和重复次数进行标准化处理后,截取相应时段的sEMG信号,计算Lempel-Ziv复杂度和Kolmogorov熵,观察它们随肌肉疲劳发展的变化规律。研究结果表明,无论是静态还是动态疲劳运动条件下,被检肌肉sEMG信号的复杂度和熵均随着运动负荷时间呈现明显的单调递减型变化。该变化可能与神经系统渐进性协调众多运动单位同步收缩的‘协同效应”有关。     

表面肌电信号的下肢痉挛信号的特征分析与识别

脑卒中患者康复治疗中会引起下肢肌肉痉挛,这种现象给患者的康复训练过程带来极大的危害,因此能够在训练过程中识别痉挛并及时中断训练具有重要的实际意义。本研究通对下肢表面肌电信号的采集,采用基于形状的模版匹配法来识别痉挛信号,并以皮尔逊相关系数来分析表征下肢痉挛信号的相关性大小。分析结果表明,通过仿真验证了模版匹配法在个人痉挛信号识别中的准确性,显示了在泛用痉挛信号识别中的可行性。     

基于表面肌电信号的人体动作识别算法研究进展

表面肌电信号(Surface Electromyography,sEMG)是通过相应肌群表面的传感器记录下来的一维时间序列非平稳生物电信号,不但反映了神经肌肉系统活动,对于反映相应动作肢体活动信息同样重要。而模式识别是肌电应用领域的基础和关键。为了在应用基于表面肌电信号模式识别中选取合适算法,本文拟对基于表面肌电信号的人体动作识别算法进行回顾分析,主要包括模糊模式识别算法、线性判别分析算法、人工神经网络算法和支持向量机算法。模糊模式识别能自适应提取模糊规则,对初始化规则不敏感,适合处理s EMG这样具有严格不重复的生物电信号;线性判别分析对数据进行降维,计算简单,但不适合大数据;人工神经网络可以同时描述训练样本输入输出的线性关系和非线性映射关系,可以解决复杂的分类问题,学习能力强;支持向量机处理小样本、非线性的高维数据优势明显,计算速度快。比较各方法的优缺点,为今后处理此类问题模式识别算法选取提供了参考和依据。     

脑卒中患者肌电信号研究

近年来,随着现代化水平的进一步加快,人们在日常生活和工作中常常会出现用脑过度现象。脑卒中的发病率也随之增加,本文主要从其发病原因及肌电信号等方面进行分析研究,为以后的临床治疗奠定基础。     

基于支持向量机的表面肌电信号和加速度融合跌倒识别方法

目的:针对老人易跌倒和跌倒过后可能产生严重后果这一现实问题,通过将表面肌电信号和加速度融合,进一步优化采用支持向量机分类器下的包含跌倒在内的几种不同动作的分类效果。方法:提出基于表面肌电和加速度信号融合的跌倒识别算法,首先采集股直肌,股内侧肌,胫骨前肌和腓肠肌的表面肌电信号以及位于腰部的三轴加速度信号作为实验数据,然后利用滑动窗口法提取表面肌电和加速度信号的均方根值,最后针对人体日常活动和跌倒的运动特征,构建了支持向量机的分类器。结果:实验数据共计320组数据,包括3种日常活动和向前跌倒,其中160组数据作为训练集,另外160组数据作为测试集。对4种动作进行识别实验,算法的准确度为93.23%、灵敏度为92.4%、特异度为100%,达到了良好的分类效果。结论:基于支持向量机的表面肌电信号和加速度融合的跌倒识别算法分类效果良好,对于老人跌倒防护具有现实意义。     

肌电信号分解及其研究现状与展望

生理学中把一个α运动神经元及其所支配的肌纤维的集合称为运动单位（ＭｏｔｏｒＵｎｉｔ）,它是神经肌肉系统的最小控制单位。α神经在神经支配或外加刺激作用下,产生一次细胞膜内外的电位变化,即神经冲动。一系列的神经冲动沿神经肌肉接头传递到所支配的肌纤维,一个...     

基于定量分析方法的动作表面肌电信号分析

介绍了非线性数据处理方法递归图法(recurrence plots,RP)及其定量分析方法(recurrence quantifi-cation analysis,RQA),并利用RP和RQA研究了动作表面肌电信号。研究发现,表面肌电信号在不同动作模式下其所对应的RP图在结构上差异明显,通过计算两通道肌电信号的RQA指标递归率,发现不同动作信号的RQA指标递归率值具有不同的聚类分布。该方法为肌电信号的动作模式分类提供了一种新的思路。     

绿化带降噪机理及模型研究进展

介绍了目前绿化带降噪机理及不同配置模式的绿化带降噪效果的研究进展。从单变量和多变量两方面总结了已有的绿化带降噪模型,普适性较差,精确性较低是目前已有模型存在的主要问题。同时,对绿化带降噪模型的研究前景进行了展望,认为可以使用非线性生态模型找出影响绿化带降噪量的主要参数,并分析主要参数与降噪量的关系,最终拟合出不同绿化带降噪量的非线性预测模型,从而指导道路绿化带的设计与建设。     

Surface electromyography as a tool to assess the responses of car passengers to lateral accelerations: Part I. Extraction of relevant muscular activities from noisy recordings

The aim of this paper is to develop a method to extract relevant activities from surface electromyography (SEMG) recordings under difficult experimental conditions with a poor signal to noise ratio. High amplitude artifacts, the QRS complex, low frequency noise and white noise significantly alter EMG characteristics. The CEM algorithm proved to be useful for segmentation of SEMG signals into high amplitude artifacts (HAA), phasic activity (PA) and background postural activity (BA) classes. This segmentation was performed on signal energy, with classes belonging to a χ² distribution. Ninety-five percent of HAA events and 96.25% of BA events were detected, and the remaining noise was then identified using AR modeling, a classification based upon the position of the coordinates of the pole of highest module. This method eliminated 91.5% of noise and misclassified only 3.3% of EMG events when applied to SEMG recorded on passengers subjected to lateral accelerations.     

Characteristics of Surface Electromyography of Forehand Smash of Badminton Players

To understand the characteristics of the forehand smash of badminton player and improve their performance, this study took eight badminton players as the subject, obtained the kinematics data through the Qualisys infrared high-speed camera, obtained the electromyography (EMG) data through the ME-6000 surface EMG test system, and compared and analyzed their forehand smash action. The results showed that the greater the angle and speed of different joints in the forehand smash was, the greater the speed and strength of hitting the ball was; the discharge amount of biceps brachii (BB) was the smallest, followed by triceps brachii (TB), flexor carpi ulnaris (FCU), anterior deltoid (AD), posterior deltoid (FD), and pectoralis major (PM), and the activation order was PM → AD → FD → BB → TB → FCU; deltoid muscle and pectoralis major muscle were the main muscle groups in the exercise, which showed the characteristic that trunk muscles drove arm muscles.     

The Kinematics and Surface Electromyography Characteristics of Round Kick of Martial Arts Athletes

In order to improve the level of athletes, modern scientific and technological means can be used to understand the characteristics and rules of movement. This study mainly analyzed the whip leg technique of Sanda athletes.Taking ten athletes as an example, the kinematics and surface electromyography(sEMG) data of them were measured, calculated and sorted out when they weredoing the action of round kick. The results showed that the movement completiontime of the first-level athletes was shorter, 0.34 ± 0.33 s. In the stage of turning hipand hitting, the angle of hip joint increased significantly. In the stage of turninghip, there was a significant difference in the angle of hip joint between differentlevels of athletes (p < 0.05), and there was no significant difference in other kinematics characteristics. In the aspect of sEMG, the duration of muscle discharge ofthe first-level athletes was shorter, but there was no significant difference in integrated electromyogram (IEMG) and root mean square (RMS). The experimentalresults reveal the importance of hip joint in the course of round kick and providesome theoretical bases for improving the level of athletes and carrying outtargeted training.     

表面肌电图在腰痛患者ODI指数和JOA评分评估中的临床应用

摘要 目的：探讨表面肌电图在腰痛患者腰椎Oswestry功能障碍指数（ODI）和日本骨科协会评估治疗分数（JOA）评估中的临床应用。方法：选择2019年6月至2020年6月我院接诊的80例腰痛患者进行研究,通过将患者按照腰部VAS评分的不同划为对照组（VAS评分≤5分）和观察组（5分＜VAS评分＜10分）,每组各40例,两组患者均接受接受常规治疗和肌电仪检测。比较治疗前后两组患者运动传导速度(MCV)、股神经的感觉传导速度(SCV)、动作电位的潜伏期、长肌力（IMS）、腰背肌后伸活动度（ROM）、ODI指数和JOA评分的变化情况。结果：治疗后,观察组运动传导速度、股神经的感觉传导速度指标水平均低于对照组,动作电位的潜伏期长于对照组（P＜0.05）;观察组长肌力、腰背肌后伸活动度指标水平均低于对照组（P＜0.05）;观察组Oswestry功能障碍指数（ODI）高于对照组,日本骨科协会评估治疗分数（JOA）评分低于对照组（P＜0.05）。结论：腰痛患者中存在着明显的表面肌电图信号改变,且随着腰痛程度的加剧,改变程度越明显,有助于评估患者病情。     

Comparative study of deep neural networks with unsupervised Noise2Noise strategy for noise reduction of optical coherence tomography images

As a powerful diagnostic tool, optical coherence tomography (OCT) has been widely used in various clinical setting. However, OCT images are susceptible to inherent speckle noise that may contaminate subtle structure information, due to low-coherence interferometric imaging procedure. Many supervised learning-based models have achieved impressive performance in reducing speckle noise of OCT images trained with a large number of noisy-clean paired OCT images, which are not commonly feasible in clinical practice. In this article, we conducted a comparative study to investigate the denoising performance of OCT images over different deep neural networks through an unsupervised Noise2Noise (N2N) strategy, which only trained with noisy OCT samples. Four representative network architectures including U-shaped model, multi-information stream model, straight-information stream model and GAN-based model were investigated on an OCT image dataset acquired from healthy human eyes. The results demonstrated all four unsupervised N2N models offered denoised OCT images with a performance comparable with that of supervised learning models, illustrating the effectiveness of unsupervised N2N models in denoising OCT images. Furthermore, U-shaped models and GAN-based models using UNet network as generator are two preferred and suitable architectures for reducing speckle noise of OCT images and preserving fine structure information of retinal layers under unsupervised N2N circumstances.