关键词： nuclear magnetic resonance   nuclear quadrupole resonance   radio frequency interference   machine learning classification  

摘要： Radio frequency interference places a major limitation on the in-situ use of unshielded nuclear quadrupole or nuclear magnetic resonance methods in industrial environments for quality control and assurance applications. In this work, we take the detection of contraband in an airport security-type application that is subject to burst mode radio frequency interference as a test case. We show that a machine learning decision tree model is ideally suited to the automated identification of interference bursts, and can be used in support of automated interference suppression algorithms. The usefulness of the data processed additionally by the new algorithm compared to traditional processing is shown in a receiver operating characteristic (ROC) analysis of a validation trial designed to mimic a security contraband detection application. The results show a highly significant increase in the area under the ROC curve from 0.580 to 0.906 for the proper identification of recovered data distorted by interfering bursts.

关键词： surface electromyography   artificial neural network   real-time   gesture recognition  

摘要： In recent years, surface electromyography (sEMG) signals have been increasingly used in pattern recognition and rehabilitation. In this paper, a real-time hand gesture recognition model using sEMG is proposed. We use an armband to acquire sEMG signals and apply a sliding window approach to segment the data in extracting features. A feedforward artificial neural network (ANN) is founded and trained by the training dataset. A test method is used in which the gesture will be recognized when recognized label times reach the threshold of activation times by the ANN classifier. In the experiment, we collected real sEMG data from twelve subjects and used a set of five gestures from each subject to evaluate our model, with an average recognition rate of 98.7% and an average response time of 227.76 ms, which is only one-third of the gesture time. Therefore, the pattern recognition system might be able to recognize a gesture before the gesture is completed.

关键词： Sensors   Bisphenols A   F   Z   stochastic sensor   water  

摘要： Three stochastic sensors based on modified nanocarbon paste (nC), nanographene paste (nGr), and reduced graphene oxide paste (rGOx) with 2, 2-diphenyl-1-picrylhydrazyl (DPPH) were proposed for the pattern recognition of bisphenols A (BPA), F (BPF), and Z (BPZ) in waste water samples. The linear concentration ranges recorded were: for BPA between 1fmol/L and 10 mu mol/L, for BPF between 10fmol/L and 1 mu mol/L, and for BPZ between 10fmol/L and 0.1mmol/L. The lowest limits of determination for each bisphenol were as following: 1fmol/L for BPA when the stochastic sensor based on nC was used, 10fmol/L for BPF when the stochastic sensors based on nC, nGr, and rGOx were used, and for BPZ 10gmol/L when the sensor based on nGr was used. The recoveries of the BPA, BPF, and BPZ were higher than 98.00% with relative standard deviation less than 0.20% (N = 10). (C) 2019 The Electrochemical Society.

关键词： Gray treefrog   communication   call segregation   cocktail party effect   acoustic interference  

摘要： Male gray treefrogs, Hyla versicolor, advertise for mates in dense assemblages characterized by high levels of noise and acoustic clutter. In pairwise interactions, males alternate pulsatile advertisement calls and so reduce call overlap to levels below that expected by chance. However, in choruses consisting of more than two males, acoustic interference increases dramatically. Moreover, males do not seem to exhibit selective attention in a way that reduces call interference among nearest neighbours. Previous research has also demonstrated that although females discriminate strongly against overlapped calls, negative effects of call overlap can be attenuated by a large angular separation between signal sources. However, call stimuli employed were identical in spectrum and so this situation differs from that likely in nature. Based on studies of 'auditory stream segregation' with other taxa, we hypothesized that realistic differences in the frequencies of overlapping calls could improve the ability of females to discern critical call features during overlap of separated call sources. We found that, although, under some circumstances, differences in call frequency may help females distinguish among neighbouring males giving temporally proximate calls, naturalistic spectral differences do not seem to help females perceptually separate the overlapping calls of neighbouring conspecific males.

关键词： load disaggregation   artificial intelligence   cognitive meters   machine learning   state machine   NILM  

摘要： A new generation of smart meters are called cognitive meters, which are essentially based on Artificial Intelligence (AI) and load disaggregation methods for Non-Intrusive Load Monitoring (NILM). Thus, modern NILM may recognize appliances connected to the grid during certain periods, while providing much more information than the traditional monthly consumption. Therefore, this article presents a new load disaggregation methodology with microscopic characteristics collected from current and voltage waveforms. Initially, the novel NILM algorithm-called the Power Signature Blob (PSB)-makes use of a state machine to detect when the appliance has been turned on or off. Then, machine learning is used to identify the appliance, for which attributes are extracted from the Conservative Power Theory (CPT), a contemporary power theory that enables comprehensive load modeling. Finally, considering simulation and experimental results, this paper shows that the new method is able to achieve 95% accuracy considering the applied data set.

关键词： 缺陷检测   图像增强   模式识别  

摘要： 目的针对当前印刷缺陷检测系统中存在的低对比度印刷缺陷检测精度不高等问题,基于HSV颜色空间,提出一种增强的低对比度印刷缺陷识别方法。方法首先,将标准样张图像与采集到的印刷图像由RGB颜色空间转换到HSV颜色空间,并提取视觉上变化敏感的亮度分量V作为待检测对象;其次,将对比度受限的局部直方图均衡(CLAHE)与数学形态学相结合,来增强显现待检测图像中的缺陷;再次,使用连通域分析方法来获取缺陷的面积、周长、离心率、长宽比和圆形度等5种特征信息,并以此建立15个特征模型;最后,构建基于PNN的印刷缺陷识别神经网络,并在Matlab中实现对低对比度印刷缺陷的识别。结果15个模型的平均耗时为475 ms,都控制在毫秒级别,满足了现代印刷缺陷检测对于实时性的要求。其中模型2的测试正确率为95%,能够识别污点等点缺陷,模型3和模型12的测试正确率为93%和93.3%,能够识别刮痕等线缺陷,模型5的测试正确率为93.1%,能够识别墨迹等面缺陷,且测试正确率高于基于BP神经网络的缺陷识别方法。结论从缺陷检测的实时性和精确性上来讲,提出的方法能够对低对比度印刷缺陷进行实时和精确的检测。

关键词： 中国大学生   网络行为   模式识别   大规模日志分析  

摘要： [目的/意义]深入挖掘和准确理解中国大学生日常网络行为模式,不仅对促进用户行为和检索领域的发展具有巨大的理论意义,而且在提升面向大学生用户的企业个性化服务与信息推荐能力方面也具有潜在的社会价值和实践意义。[方法/过程]提出一种基于大规模日志分析的大学生用户行为模式识别新方法,该方法包括一种基于深度学习和文本分析技术的半监督学习算法"MaxMatching"以及混合两种特征熵(香农熵与真实熵)的聚类模型。[结果/结论]实证结果表明本方法不仅在算法和结果解释上具有一定的优势,而且能从网络使用能力、访问时序性和主题倾向性三方面归纳与呈现中国大学生网络行为全方位模式。该方法和结论有效地拓展了信息检索领域查询项的语义化理解方面的方法体系,也为企业提升面向大学生用户的个性化信息推荐服务提供一定的参考和可行性意见。

关键词： 振动模式识别   仿真分析   新型传感光纤  

摘要： 模式识别是人工智能的重要组成部分,通过模式识别可以事先预防和减少恶性事故的发生,本文通过对新型传感光纤传输的理论分析和性能的仿真,给出光纤受到振动时,模式场的变化及结合振动的类型,给出仿真分析结果。

关键词： 绞股蓝   指纹图谱   液质联用   聚类分析   主成分分析  

摘要： 目的采用LC-MS建立绞股蓝药材的指纹图谱,鉴定其主要色谱峰,并结合变量模式识别方法判别不同产地绞股蓝药材。方法采用Inertsil ODS-SP(250mm×4.6mm,5μm)C18色谱柱,以乙腈-0.1%甲酸水为流动相梯度洗脱,检测波长203nm,使用电喷雾离子源(ESI),负离子模式下采集数据。通过聚类分析、主成分分析和对应分析对21批绞股蓝药材进行分析。结果建立了绞股蓝药材指纹图谱,标定出9个共有峰与10个特征峰,并根据质谱信息及文献对照对其中16个色谱峰进行定性分析。并结合多变量分析区分了不同产地绞股蓝,为绞股蓝的质量控制及综合评价奠定了基础。结论液质联用-模式识别方法可用于不同产地绞股蓝的快速鉴别,也可对其特征成分进行靶向分离。

关键词： 图像处理   模式识别   特征提取   人脸识别   隐式低秩表示  

摘要： 隐式低秩表示(LatLRR)作为经典的无监督特征提取算法已应用于模式识别领域。然而该算法得到的特征维数无法降低,且由于算法分别学习2个低秩矩阵,因此无法保证整体最优;另外,算法忽略了样本在学习过程中存在的残差。为解决这些问题,提出了一种基于隐式低秩表示的联合投影学习算法。使用投影矩阵和恢复矩阵近似地表示隐式低秩表示中的投影矩阵,使算法在降维的同时可提取判别特征;联合学习投影矩阵、恢复矩阵和低秩矩阵,矩阵间相互提升,从获得的投影中可以提取出更多的判别特征,同时在算法模型中约束样本在投影学习中存在的残差;最后采用交替迭代方法求解该模型。在多个数据集上进行实验,结果说明算法在有效降维的同时能进一步提高判别能力。