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。     

基于支持向量机的室颤信号检测算法

目的:实现室颤信号与非室颤信号的分类,进而实现室颤信号的检测。方法:本文引入了一种基于支持向量机(Support Vec-tor Machine,SVM)和改进的越限区间算法(TCI)的新算法,其中支持向量机在处理分类和模式识别等问题中具有很大的优势。该算法采用4s的滑动窗技术,并利用改进后的越限区间算法(Threshold Crossing Interval,TCI)方法提取心电信号的特征。新算法的实现如下:在每一滑动窗内采用改进的后的绝对值阈值,计算中间2s内的平均越限间隔值。并以此TCI值作为特征参数,输入一个预先设计好的二分类支持向量机中,从而实现分类。结果:成功实现了室颤信号的检测,通过计算该方法的灵敏度、精确度、预测性和准确度且与其他方法相比较,可知此新算法总体可靠性优于其他方法。结论:该算法能够实现室颤信号的实时监测,且简单易行,易于实现,较适合实时的心电监测以及除颤仪器。     

支持向量机在PPⅡ二级结构预测中的应用

PPⅡ二级结构是一种稀有的蛋白质结构类型。目前使用机器学习方法预测此二级结构的工作还比较少见。引入一种新的方法———支持向量机 (SVM)来预测PPII二级结构 ,并与神经网络方法进行了比较 ,结果表明 ,SVM方法在预测PPII结构上表现良好 ,预测精度达到 76 .5 2 %。     

苏云金杆菌杀虫晶体蛋白活性预测的支持向量机模型

藉均匀设计(UD)方法,构建了苏云金杆菌(Bt)杀虫晶体蛋白氨基酸组成特征与其杀虫活性之间关系的支持向量机(SVM)模型。当惩罚系数为0·01、epsilon值为0·2、gamma值为0·05、域值为0·5时,该模型对Bt杀虫晶体蛋白杀虫活性的预测平均准确率达73%。     

用稳定性同位素技术确定高寒草甸生态系统中动物营养级模型

根据稳定性同位素技术原理建立了高寒草甸生态系统中动物的营养级模型 (3)、 (7)和 (9)式。 3个模型分别描述了每种食物资源对动物的贡献大小 (PCV)、食物资源 (Ai)占取食动物 (P)的整个食物的比例 (PAiP)、动物在高寒草甸生态系统中的营养级 (TLc) :PCVAi =cos (ΔαPAi)ΖPAi(3)PAiP =PCVAi∑i=1PCVAi× % (7)TLc =1 (αc-αTL1) /Δαcd (9)式中 ,ΔαPAi为捕食者P与食物Ai 的取食角 ,ΖPAi为捕食向量与食物向量之间的欧氏距离 ,αc 是消费者的向量角 ,αTL1是第一营级的向量角 ,利用系数Δαcd是消费者与食物向量角之差 (为一常数 )。同时 ,给出了判断高寒草甸两个物种之间捕食或营养关系模型 (ΖS1 S2 ) :当cos (Δα) /PCVmin≤ΖS1 S2 ≤cos (Δα) /PCVmax时存在捕食关系 ,并为上下级营养关系 ;当ΖS1 S2 cos (Δα) /PCVmax时 ,不存在捕食关系 ,前式为同一营养级 ,后式为相隔一个至几个营养级。模型 (9)式得到的结果与张晓爱等 (1999)报道一致。     

基于去趋势波动分析和支持向量机的癫痫脑电分类

癫痫脑电的自动检测与分类对于癫痫患者的诊断治疗具有重要意义.本文提出了一种基于去趋势波动分析(detrended fluctuation analysis,DFA)、波动指数和支持向量机(support vectormachine,SVM)的癫痫脑电分类方法.首先采用DFA方法计算脑电的标度指数,然后对脑电进行4～8 ...     

贵州喀斯特地区辣椒镉的累积特性及土壤风险阈值研究

辣椒(Capsicum annuum L.)是我国西南喀斯特重金属地质高背景区的特色农作物, 属茄科, 具有一定镉(Cadmium)富集能力。为明确贵州主产区辣椒Cd的富集状况, 探究辣椒质量安全的土壤风险阈值, 保障农产品安全生产。采集贵州省9个县区辣椒及土壤样品105组, 利用ICP-MS检测Cd含量, 研究土壤Cd分布, 辣椒富集系数, 并根据土壤pH建立回归模型推导基于辣椒质量安全的土壤Cd风险阈值。结果表明, 有4个研究区土壤均值超过贵州省土壤Cd背景值(0.659 mg·kg-1), 超出率分别是: WS(25.8%)、HZ(42.2%)、LD(146%)、WN(439%); 18.1%的辣椒超过国家食品安全Cd限值标准(GB2762—2017, 0.05 mg·kg-1), 点位超标率为LD(20.8%)、TZ(7.69%)和WN(100%); 辣椒Cd的生物富集系数(BCF)范围是0.019至0.108, 均值0.046。由回归模型推算出酸性土(pH<6.5)、中性土(6.5≤pH≤7.5)、石灰性土(pH>7.5)中Cd风险阈值分别为1.00、1.26和2.50 mg·kg-1, 分别为国家农用地土壤污染风险筛选值(GB15618—2018)的3.33、4.21、4.17倍, 阈值准确率达91.1%; 有效态Cd风险阈值为0.071、0.017和0.005 mg·kg-1, 阈值准确率达86.1%。地质高背景区农田土壤Cd超标严重, 但辣椒的生物富集系数相对较低, 基于辣椒质量安全的土壤Cd风险阈值远高于国家标准。     

两种过滤特征基因选择算法的有效性研究

对基因表达谱进行特征基因选择不仅能改善疾病分类方法的效能,而且为寻找与疾病相关的特征基因提供新的途径．通过比较用调整p值的t检验、非参数评分两种特征基因选择算法后和未进行选择时支持向量机(SVM)分类器的分类性能、支持向量(SV)的吻合度、错分样本ID的吻合度和对样本均匀翻倍后的稳定性．结果发现：特征选择后线性、核函数为二阶多项式和径向基的SVM分类性能明显提高;特征选择前后的SV及错分样本ID的吻合度均较高;SVM的稳定性较好．由此得出结论：这两种特征选择算法具有一定的有效性．     

基于PSO-LSSVR模型的土壤肥力评价

通过引入粒子群算法(PSO)和最小二乘支持向量机(LSSVR),提出基于PSO-LSSVR的土壤肥力评价模型。选取有机质、全氮速效磷、速效钾、阳离子交换量、酸碱度、容重、黏粒、水稳性团聚体和分散率等10种评价指标,以吉林省黑地为例,建立土壤肥力评价模型。同时与物元可拓法、普通SVM模型的评价结果进行比较;3种方法的多数样本评价结果基本一致,对于样本2、样本13,PSO-LSSVR模型分别定为Ⅳ级、Ⅲ级,符合实际情况;表明PSO-LSSVR是一种适用且能准确反映土壤特性的土壤肥力评价模型。     

奇异数据筛选法在玉米籽粒蛋白质近红外光谱检测中的应用

相对于传统生化测定方法,基于近红外光谱(Near infrared spectroscopy,NIRS)玉米籽粒蛋白质含量检测是一种快速、非破坏、且适用于多组分同时检测的新方法。但在建模过程中,由于奇异数据(异常值)的存在会影响近红外光谱模型的预测精度和稳定性,我们采用奇异数据筛选法剔除了玉米籽粒近红外光谱中的奇异数据并建立了玉米籽粒蛋白质含量的偏最小二乘支持向量机(Least squares support vector machine,LS-SVM)模型。本文分别采用杠杆值法(Leverage)、半数重采样法(Resampling by Half-Mean,RHM)和蒙特卡洛采样法(Monte-Carlo Sampling,MCS)剔除了玉米籽粒蛋白质光谱数据中的奇异数据并对模型结果进行比较。在剔除奇异数据的基础上,采用偏最小二乘回归法(Partial least squares regression,PLSR)提取主成分,并基于小生境蚁群算法(Niche ant colony algorithm,NACA)优化偏最小二乘支持向量机(LS-SVM)模型参数(γ和σ2),建立基于LS-SVM的玉米籽粒蛋白质定量分析模型。结果表明,采用3种奇异数据筛选法剔除奇异数据后所建LS-SVM模型的预测结果都优于采用原光谱数据所建模型,相比较而言,蒙特卡洛采样法为基于近红外光谱检测玉米籽粒蛋白质的最佳奇异数据筛选法。     

Detecting Falls as Novelties in Acceleration Patterns Acquired with Smartphones

Despite being a major public health problem, falls in the elderly cannot be detected efficiently yet. Many studies have used acceleration as the main input to discriminate between falls and activities of daily living (ADL). In recent years, there has been an increasing interest in using smartphones for fall detection. The most promising results have been obtained by supervised Machine Learning algorithms. However, a drawback of these approaches is that they rely on falls simulated by young or mature people, which might not represent every possible fall situation and might be different from older people''s falls. Thus, we propose to tackle the problem of fall detection by applying a kind of novelty detection methods which rely only on true ADL. In this way, a fall is any abnormal movement with respect to ADL. A system based on these methods could easily adapt itself to new situations since new ADL could be recorded continuously and the system could be re-trained on the fly. The goal of this work is to explore the use of such novelty detectors by selecting one of them and by comparing it with a state-of-the-art traditional supervised method under different conditions. The data sets we have collected were recorded with smartphones. Ten volunteers simulated eight type of falls, whereas ADL were recorded while they carried the phone in their real life. Even though we have not collected data from the elderly, the data sets were suitable to check the adaptability of novelty detectors. They have been made publicly available to improve the reproducibility of our results. We have studied several novelty detection methods, selecting the nearest neighbour-based technique (NN) as the most suitable. Then, we have compared NN with the Support Vector Machine (SVM). In most situations a generic SVM outperformed an adapted NN.     

How Much Gravity Is Needed to Establish the Perceptual Upright?

Might the gravity levels found on other planets and on the moon be sufficient to provide an adequate perception of upright for astronauts? Can the amount of gravity required be predicted from the physiological threshold for linear acceleration? The perception of upright is determined not only by gravity but also visual information when available and assumptions about the orientation of the body. Here, we used a human centrifuge to simulate gravity levels from zero to earth gravity along the long-axis of the body and measured observers'' perception of upright using the Oriented Character Recognition Test (OCHART) with and without visual cues arranged to indicate a direction of gravity that differed from the body''s long axis. This procedure allowed us to assess the relative contribution of the added gravity in determining the perceptual upright. Control experiments off the centrifuge allowed us to measure the relative contributions of normal gravity, vision, and body orientation for each participant. We found that the influence of 1 g in determining the perceptual upright did not depend on whether the acceleration was created by lying on the centrifuge or by normal gravity. The 50% threshold for centrifuge-simulated gravity''s ability to influence the perceptual upright was at around 0.15 g, close to the level of moon gravity but much higher than the threshold for detecting linear acceleration along the long axis of the body. This observation may partially explain the instability of moonwalkers but is good news for future missions to Mars.     

Application of motion analysis system in pre-impact fall detection

The purpose of this study is to investigate unique features of body segments in fall and activities of daily living (ADL) to make automatic detection of fall in its descending phase before the impact. Thus, fall-related injuries can be prevented or reduced by deploying feedback systems before the impact. In this study, the authors propose the following hypothesis: (1) thigh segment normally does not go beyond certain threshold angle to forward and sideways directions in ADL and (2) even if it does, the angular characteristics measured at torso and thigh differ from one another in ADL whereas in the case of fall, they become congruent. These two factors can be used to distinguish fall from ADL in its inception. Vicon 3-D motion analysis system was used in this study. High level of correlation between thigh and torso segments (corr > 0.99) was found for fall activities and low correlation coefficients (mean corr for lateral movements is 0.2338 and for sagittal movements is -0.665) were observed in ADL. By applying the hypothesis, all simulated falls could be detected with no false alarms and around 700ms lead-time before the impact was achieved in pre-impact fall detection. It is the longest lead-time obtained so far in pre-impact fall detection.     

Detection-threshold of 50-Hz electric fields by human subjects

The threshold intensity for detection of an AC electric field was studied in human subjects at several different temperatures and humidities. The dorsum and palm of the hand were exposed to fields, representing hairy and hairless skin, in order to clarify whether hair movement is critical for field detection. Experiments were carried out on human subjects (seven men and four women) during hot humid weather of July–August and dry cool air of October–November. Threshold values obtained in the summer were 30–65 kV/m for the hairy skin on the dorsum of the hand, while for the hairless skin on the palm the threshold was > 115 kV/m (highest field available due to limitations of the power supply). During the fall, the threshold was much higher than during the summer. We sought possible reasons for the difference and found that humidity was the main factor. Relative permittivity of woman's hair was then estimated by measuring capacitance of the hairs under dry (35% RH) and wet (85% RH) conditions at 20 °C. The values of relative permittivity obtained under these two conditions differed by several times the average. The differences in detection thresholds may be attributable to the different relative permittivities of the hairs under dry and wet conditions.     

Internally generated preactivation of single neurons in human medial frontal cortex predicts volition

Understanding how self-initiated behavior is encoded by neuronal circuits in the human brain remains elusive. We recorded the activity of 1019 neurons while twelve subjects performed self-initiated finger movement. We report progressive neuronal recruitment over ～1500 ms before subjects report making the decision to move. We observed progressive increase or decrease in neuronal firing rate, particularly in the supplementary motor area (SMA), as the reported time of decision was approached. A population of 256 SMA neurons is sufficient to predict in single trials the impending decision to move with accuracy greater than 80% already 700 ms prior to subjects' awareness. Furthermore, we predict, with a precision of a few hundred ms, the actual time point of this voluntary decision to move. We implement a computational model whereby volition emerges once a change in internally generated firing rate of neuronal assemblies crosses a threshold.     

Systolic Peak Detection in Acceleration Photoplethysmograms Measured from Emergency Responders in Tropical Conditions

Photoplethysmogram (PPG) monitoring is not only essential for critically ill patients in hospitals or at home, but also for those undergoing exercise testing. However, processing PPG signals measured after exercise is challenging, especially if the environment is hot and humid. In this paper, we propose a novel algorithm that can detect systolic peaks under challenging conditions, as in the case of emergency responders in tropical conditions. Accurate systolic-peak detection is an important first step for the analysis of heart rate variability. Algorithms based on local maxima-minima, first-derivative, and slope sum are evaluated, and a new algorithm is introduced to improve the detection rate. With 40 healthy subjects, the new algorithm demonstrates the highest overall detection accuracy (99.84% sensitivity, 99.89% positive predictivity). Existing algorithms, such as Billauer''s, Li''s and Zong''s, have comparable although lower accuracy. However, the proposed algorithm presents an advantage for real-time applications by avoiding human intervention in threshold determination. For best performance, we show that a combination of two event-related moving averages with an offset threshold has an advantage in detecting systolic peaks, even in heat-stressed PPG signals.     

Bending it like Beckham: how to visually fool the goalkeeper

Background

As bending free-kicks becomes the norm in modern day soccer, implications for goalkeepers have largely been ignored. Although it has been reported that poor sensitivity to visual acceleration makes it harder for expert goalkeepers to perceptually judge where the curved free-kicks will cross the goal line, it is unknown how this affects the goalkeeper''s actual movements.

Methodology/Principal Findings

Here, an in-depth analysis of goalkeepers'' hand movements in immersive, interactive virtual reality shows that they do not fully account for spin-induced lateral ball acceleration. Hand movements were found to be biased in the direction of initial ball heading, and for curved free-kicks this resulted in biases in a direction opposite to those necessary to save the free-kick. These movement errors result in less time to cover a now greater distance to stop the ball entering the goal. These and other details of the interceptive behaviour are explained using a simple mathematical model which shows how the goalkeeper controls his movements online with respect to the ball''s current heading direction. Furthermore our results and model suggest how visual landmarks, such as the goalposts in this instance, may constrain the extent of the movement biases.

Conclusions

While it has previously been shown that humans can internalize the effects of gravitational acceleration, these results show that it is much more difficult for goalkeepers to account for spin-induced visual acceleration, which varies from situation to situation. The limited sensitivity of the human visual system for detecting acceleration, suggests that curved free-kicks are an important goal-scoring opportunity in the game of soccer.     

Biomechanical study of the programming of anticipatory postural adjustments associated with voluntary movement

The present research concerns anticipatory postural adjustments (APA), with the purpose of determining whether they are preprogrammed and of specifying their biomechanical finality. The experimental situation allowed us to distinguish between the voluntary movement itself (an upper limb elevation) and the postural adjustments associated with it. To this aim, the upper limb kinematics, evaluated from an accelerometer fixed at wrist level, were compared to the whole body dynamics, recorded by means of a force platform. Movements, executed in series of five, were studied according to three conditions: bilateral flexions (BF) and unilateral flexions (UF), with (IUF) and without (OUF) an additional inertia, of the stretched upper limb(s). Six right handed adults were tested twice. Results showed that the ground reaction resultant forces as well as the ground reaction resultant moment about the vertical axis presented reproducible variations before and after the onset of upper limb acceleration. The biomechanical organization of APA corresponded, for the three experimental conditions, to an upward and forward acceleration of the body center of gravity, and also, for UF, to a resultant moment directed towards the contralateral side. The duration of APA varied with the characteristics of the forthcoming voluntary movement, increasing significantly from BF to OUF and from OUF to IUF. It is concluded that APA correspond to dynamic phenomena which are centrally preprogrammed. The inertia forces associated with APA may, when the time comes, balance the inertia forces due to the movement of the mobile limb therefore counteracting the disturbance to postural equilibrium.(ABSTRACT TRUNCATED AT 250 WORDS)     

Polar body mutation load analysis in a patient with A3243G tRNALeu(UUR) point mutation

Diseases associated with point mutations in the mitochondrial DNA (mtDNA) are maternally inherited. We evaluated whether pre-implantation genetic diagnosis, based on polar body mutation load detection could be used to distinguish healthy from affected oocytes. Restriction Fragment Length Polymorphism (RFLP) analysis was used and validated, to determine A3243G tRNA(Leu(UUR)) mutation load in metaphase II oocytes and their respective first polar bodies. The results of this study show for the first time that the mutation load measured in the polar bodies correlates well with the mutation load in the respective oocytes. Therefore, human polar body analysis can be used as diagnostic tool to prevent transmission of mitochondrial disorders.