基于SVM的脑功能分类与识别方法研究

目的探讨SVM分类器用于脑功能识别的可行性、有效性与优越性,为脑电信号处理及功能识别提供一种新的途径和参考。方法对400组实测的正常人在睁眼和闭眼两种状态下的脑电信号,选取四种核函数分别构造四种SVM分类器对上述两种状态下的脑功能进行分类识别,从不同角度深入分析和比较讨论了由四种核函数构造的SVM分类器性能,并提出了脑电信号特征参数从低维到高维的组合变换新方法。结果由RBF核函数构造的SVM分类器最为适合脑功能的分类识别,正识率最高可达96%。结论支持向量机的方法用于脑电信号处理及功能模式识别是可行的、有效的、并初步表现出了优越的性能。     

基于三维谱图混沌特征的GIS局部放电识别

气体绝缘电器(GIS)局部放电(PD)的故障诊断对于GIS的运行状态评估有着重要意义,传统模式识别方法局限于对描述PD谱图形态分布方面的特征进行分析与识别,缺乏对PD特征更全面、更深刻、更本质的分析研究,导致出现对某些类型放电识别率低等问题。针对这些问题,本文提出了一种基于混沌理论的GIS PD识别方法,连续采集100个工频周期的PD信号构成一个f-v-n三维谱图样本F,以矩阵F的一列作为一个信号序列进行混沌分析,即计算对应同一相位信号序列的最大Lyapunov指数,获取36个最大Lyapunov指数在不同相位区间的分布特征作为不同相位下的局部放电混沌特征。实验结果表明,提取的混沌特征可实现对PD本质的深入挖掘,整体识别效果较好,特别是对于传统的统计特征识别方法难以区分的气隙类缺陷识别率很高,可作为统计特征识别方法的辅助方法加入到识别系统中,进一步提高识别准确率。     

用谐波小波包变换法提取GIS局部放电信号多尺度特征参数

超高频(UHF)法在GIS局部放电(PD)检测中已得到了广泛应用，UHF PD信号的特征提取对准确识别GIS内部绝缘缺陷类型和指导检修工作具有重要意义，但目前仍然缺乏有效的特征提取方法。为此，本文利用谐波小波具有严格盒形频谱的优点，提出一种提取UHF PD特征信息的谐波小波包变换(HWPT)方法，对实验室获取的4种典型放电模型产生的UHF PD信号，采用HWPT进行多尺度分解，以克服实小波包分解子带间存在频谱混叠和能量泄漏的缺陷，利用UHF PD信号在不同尺度能量和复杂度的差异，提取多尺度能量和多尺度样本熵参数作为模式识别的特征量，更加精确地描述了UHF PD信号的时频域信息。最后利用支持向量机分类识别的结果表明，该方法可以取得比实小波包更好的识别效果，多尺度能量和多尺度样本熵特征参数均能有效识别4种绝缘缺陷。     

基于MR对比增强T1WI纹理分析鉴别泪腺淋巴瘤与泪腺炎性假瘤

目的 观察基于MR对比增强T1WI （CE-T1WI）纹理分析鉴别诊断泪腺淋巴瘤与泪腺炎性假瘤的价值。方法 回顾性分析经病理证实的21例泪腺淋巴瘤（淋巴瘤组）和25例泪腺炎性假瘤（炎性假瘤组）的眼眶MRI表现,基于CE-T1WI提取病灶直方图、灰度共生矩阵、灰度游程矩阵、绝对梯度、自回归模型和小波变换6种共279个纹理特征参数,采用组间比较、组内相关系数（ICC）及最小绝对收缩和选择算子（LASSO）回归筛选最佳纹理特征,建立核函数分别为线性核（LK）、多项式核（PK）和径向基函数核（RBFK）的支持向量机（SVM）分类模型,筛选最优核函数。针对最佳纹理特征及组间差异有统计学意义的MRI表现,以最优核函数建立联合模型;并以受试者工作特征（ROC）曲线评估各模型鉴别诊断泪腺淋巴瘤与泪腺炎性假瘤的效能。结果 相比炎性假瘤组,淋巴瘤组病灶边界更清晰、强化更均匀（P均<0.01）,组间其余MRI表现差异无统计学意义（P均>0.05）。共201个纹理特征组间差异有统计学意义,经筛选10个最佳纹理特征用于建立SVM分类模型,其中PK为最优核函数,相应SVM分类模型鉴别诊断泪腺淋巴瘤与泪腺炎性假瘤的效能最佳,其敏感度、特异度、准确率及曲线下面积（AUC）分别为90.47%、88.00%、89.13%及0.93,联合模型分别为95.23%、92.00%、93.47%及0.96;联合模型与最优SVM分类模型的AUC差异无统计学意义（P=0.33）。结论 基于MR CE-T1WI纹理分析可有效鉴别泪腺淋巴瘤与泪腺炎性假瘤。     

基于多变量模式分析的飞行学员脑功能连接的识别研究

目的 基于多变量模式分析（multivariate pattern analysis, MVPA）对飞行学员和健康的普通人的大脑功能连接进行有效识别。材料与方法 采集了40名已经取得执照的飞行专业在校学生与39名地面专业在校学生的功能磁共振数据。通过网络功能连接分析得到功能连接矩阵作为特征,分别通过最小绝对收缩选择算子（least absolute shrinkage and selection operator,LASSO）算法与独立样本t检验方法对特征降维。使用不同核函数的支持向量机（support vector machine, SVM）进行训练和预测,使用留一交叉验证法进行模型性能评估,最终根据训练后SVM模型中的权重定位对应脑区之间的功能连接。结果 使用LASSO特征筛选的线性（linear）核SVM模型准确率为81.82%,敏感度82.05%,特异度81.58%,曲线下面积（area under the curve, AUC）为0.88。核函数对模型准确率的影响不大。模型中右侧中央旁小叶、双侧中央后回、双侧顶下缘角回、右侧梭状回、左侧眶部额中回、左侧顶上回、右侧眶部额下回有...     

局部放电UHF脉冲的时频特征提取与聚类分析

制作了4种人工缺陷模型模拟典型的局部放电源，并进行局部放电试验采集UHF脉冲信号。引入S变换(ST)对局部放电的UHF脉冲进行时频分析，探索不同放电源脉冲的聚类分离。算法首先对UHF脉冲进行S变换，并采用非负矩阵分解(NMF)对S变换幅值矩阵进行分解得到频域基向量和时域位置向量，从中提取尖锐度、导数平方和、信息熵以及稀疏度等特征参量，构造出能充分反映局部放电时频信息的特征空间，最后利用模糊C均值算法对提取的特征向量进行聚类得到放电源脉冲的聚类结果。对试验数据的分析结果表明，提取的ST时频特征能够有效实现不同局部放电源脉冲的聚类，当NMF参数r=2时，10维时频特征能够取得最高为90.33%的聚类正确率；与常用的Wigner-Ville分布(WVD)相比，ST具有更好的聚类效果；当存在复杂的多重信号折反射时，本文提出的时频特征聚类结果较差，需要进行进一步的研究。     

基于机器学习和差值校验的识别混淆样本方法的建立及评价

目的 利用患者历史对比数据,建立基于机器学习中SVM算法识别临床混淆样本的方法,并通过与RCV方法比较,验证该方法的临床有效性。方法 从北京朝阳医院采集的22项血液常规检测结果,经过数据清洗过滤,筛选合格的患者只保留两次检验结果,两次结果按所有项目对应计算dalta值,用同一患者与不同患者的delta值分别制作正配样本与错配样本。dalta值类型分绝对值与相对值两种。采用SVM分类算法实现用多项目识别两种样本,与基于单项目识别的RCV方法作比较。探讨利用SVM算法识别混淆样本技术在临床的可行性。结果 SVM算法通过22项指标识别正配和错配两种样本,经过网格搜索法寻取最佳参数,精确度达到了0.92。基于统计学的RCV方法在不同项目下效果不同,其中MCH在绝对值delta下的准确度最高,为0.8145。结论 基于多项目的SVM算法在混淆样本的识别上准确度更高,性能优于传统方法,适合在临床上普及应用。     

基于机器学习的组合模型在预测乳腺癌新辅助化疗疗效中的价值

目的 探究基于机器学习组合模型的影像组学在预测肿块型乳腺癌新辅助化疗(neoadjuvant chemotherapy,NAC)疗效中的价值。材料与方法 回顾性分析2018年1月到2021年10月中国人民解放军总医院第五医学中心的97例接受NAC治疗且经组织病理学证实的肿块型乳腺癌患者的临床和影像资料。基于实体瘤疗效评定标准(Response Evaluation Criteria in Solid Tumors,RECIST),将患者分为有效组和无效组,基于治疗前的动态对比增强MRI(dynamic contrast-enhanced MRI,DCE-MRI)减影第一期图像上提取的影像组学特征,引入高通或低通小波滤波器和不同参数的拉普拉斯-高斯滤波器对原始MR图像进行预处理。采用基于单变量分析和多变量分析的特征选择方法进行特征筛选,单变量分析包括F检验、卡方检验和互信息;多变量分析采用最小绝对收缩和选择算子(least absolute shrinkage and selection operator,LASSO);采用支持向量机(support vector machine,SVM...     

基于DCE图像优化胶质瘤自动分级纹理模型

目的探讨不同纹理模型和灰阶对基于动态对比增强磁共振图像(dynamic contrast enhancement magnetic resonance imaging,DCE-MRI)的支持向量机的胶质瘤自动分级影响。材料与方法收集经磁共振扫描且经病理证实为胶质瘤Ⅱ、Ⅲ、Ⅳ级的患者共117例,计算DCE-MRI图像血流动力学参数(NordicICE 4.0),利用不同纹理模型和灰阶提取参数图肿瘤区域相应纹理特征。支持向量机递归特征消除算法选择特征后,输入线性SVM对胶质瘤级别进行分类并使用留一法交叉验证。分类结果使用Graphpad Prism 6统计软件分析。结果灰阶对分类效能的影响差异无统计学意义(P=0.1589),纹理模型对分类效能的影响差异存在统计学意义(P0.0001)。在使用灰度共生矩阵(gray-level cooccurrence matrix,GLCM)提取纹理特征并且灰阶为32和256时,分别选取前22个和前17个特征所得分类正确率最高(正确率=0.79)。结论基于DCE图像纹理对支持向量机胶质瘤分级中,纹理模型GLCM结合特征选择是胶质瘤分级的最优方案,并推荐在后期研究中使用。     

基于高阶神经网络的肌电信号识别方法的改进

目的;提高假肢分类训练的速度和准确率。方法：采用一种高效率的高阶神经网络－－Ｐｉ－Ｓｉｇｍａ网络,并针对肌电信号的非平稳特性,对用小波变换提取的表面肌电特征进行分类。结果：效率大大提高,而且在训练速度提高的同时,并不其分类的准确性。结论。与传统识别方法相比,Ｐｉ－Ｓｉｇｍａ网络的肌电信号识别方法训练速度快,精确度高,生好是一种具有肌电信号识别方法。     

Unsupervised analysis of fMRI data using kernel canonical correlation

We introduce a new unsupervised fMRI analysis method based on kernel canonical correlation analysis which differs from the class of supervised learning methods (e.g., the support vector machine) that are increasingly being employed in fMRI data analysis. Whereas SVM associates properties of the imaging data with simple specific categorical labels (e.g., -1, 1 indicating experimental conditions 1 and 2), KCCA replaces these simple labels with a label vector for each stimulus containing details of the features of that stimulus. We have compared KCCA and SVM analyses of an fMRI data set involving responses to emotionally salient stimuli. This involved first training the algorithm (SVM, KCCA) on a subset of fMRI data and the corresponding labels/label vectors (of pleasant and unpleasant), then testing the algorithms on data withheld from the original training phase. The classification accuracies of SVM and KCCA proved to be very similar. However, the most important result arising form this study is the KCCA is able to extract some regions that SVM also identifies as the most important in task discrimination and these are located manly in the visual cortex. The results of the KCCA were achieved blind to the categorical task labels. Instead, the stimulus category is effectively derived from the vector of image features.     

Utilizing temporal information in fMRI decoding: classifier using kernel regression methods

This paper describes a general kernel regression approach to predict experimental conditions from activity patterns acquired with functional magnetic resonance image (fMRI). The standard approach is to use classifiers that predict conditions from activity patterns. Our approach involves training different regression machines for each experimental condition, so that a predicted temporal profile is computed for each condition. A decision function is then used to classify the responses from the testing volumes into the corresponding category, by comparing the predicted temporal profile elicited by each event, against a canonical hemodynamic response function. This approach utilizes the temporal information in the fMRI signal and maintains more training samples in order to improve the classification accuracy over an existing strategy. This paper also introduces efficient techniques of temporal compaction, which operate directly on kernel matrices for kernel classification algorithms such as the support vector machine (SVM). Temporal compacting can convert the kernel computed from each fMRI volume directly into the kernel computed from beta-maps, average of volumes or spatial-temporal kernel. The proposed method was applied to three different datasets. The first one is a block-design experiment with three conditions of image stimuli. The method outperformed the SVM classifiers of three different types of temporal compaction in single-subject leave-one-block-out cross-validation. Our method achieved 100% classification accuracy for six of the subjects and an average of 94% accuracy across all 16 subjects, exceeding the best SVM classification result, which was 83% accuracy (p=0.008). The second dataset is also a block-design experiment with two conditions of visual attention (left or right). Our method yielded 96% accuracy and SVM yielded 92% (p=0.005). The third dataset is from a fast event-related experiment with two categories of visual objects. Our method achieved 77% accuracy, compared with 72% using SVM (p=0.0006).     

Scene classification via latent Dirichlet allocation using a hybrid generative/discriminative strategy for high spatial resolution remote sensing imagery

In order to capture the high-level concepts in high spatial resolution (HSR) remote sensing imagery, scene classification based on a latent Dirichlet allocation (LDA) model, a generative topic model, is a practical method to bridge the semantic gaps between the low-level features and the high-level concepts of HSR imagery. In the previous work, LDA has been considered as a scene classifier, namely C-LDA, and multiple LDA models for each scene class are built separately, where the scene class is determined by a maximum likelihood rule. The C-LDA strategy disregards the correlations between the generative topic spaces of the different scene classes. In this letter, two novel strategies of scene classification based on LDA are proposed to consider the correlations between the generative topic spaces of the different scene classes by sharing the topic spaces for all the scene classes. One of the proposed strategies utilizes LDA as part of the classifier, namely P-LDA, which generates the topic space from all the training images. A discriminative classifier (e.g., support vector machine, SVM) is also employed as the other classification part of P-LDA. The other proposed strategy employs LDA as the topic feature extractor, namely F-LDA, which generates the topic space from all the training and test images, and utilizes a discriminative classifier to classify the topic features. The experimental results using aerial orthophotographs show that the performances of the two proposed strategies for scene classification based on LDA are better than the traditional C-LDA method.     

Rapid detection of hysteromyoma and cervical cancer based on serum surface-enhanced Raman spectroscopy and a support vector machine

In this study, we investigated the feasibility of using surface-enhanced Raman spectroscopy (SERS) combined with a support vector machine (SVM) algorithm to discriminate hysteromyoma and cervical cancer from healthy volunteers rapidly. SERS spectra of serum samples were recorded from 30 hysteromyoma patients, 36 cervical cancer patients as well as 30 healthy subjects. SVM was used to establish the classification models, and three types of kernel functions, namely linear, polynomial, and Gaussian radial basis function (RBF), were utilized for comparison. When the polynomial kernel function was employed, the overall diagnostic accuracy for classifying the three groups could achieve 86.5%. In addition, when the optimal kernel function was selected, the diagnostic accuracy for identifying healthy versus hysteromyoma, healthy versus cervical cancer, and hysteromyoma versus cervical cancer reached 98.3%, 93.9%, and 90.9%, respectively. The current results indicate that serum SERS technology, together with the SVM algorithm, is expected to become a clinical tool for rapid screening of hysteromyoma and cervical cancer.     

Scalable histopathological image analysis via supervised hashing with multiple features

Histopathology is crucial to diagnosis of cancer, yet its interpretation is tedious and challenging. To facilitate this procedure, content-based image retrieval methods have been developed as case-based reasoning tools. Especially, with the rapid growth of digital histopathology, hashing-based retrieval approaches are gaining popularity due to their exceptional efficiency and scalability. Nevertheless, few hashing-based histopathological image analysis methods perform feature fusion, despite the fact that it is a common practice to improve image retrieval performance. In response, we exploit joint kernel-based supervised hashing (JKSH) to integrate complementary features in a hashing framework. Specifically, hashing functions are designed based on linearly combined kernel functions associated with individual features. Supervised information is incorporated to bridge the semantic gap between low-level features and high-level diagnosis. An alternating optimization method is utilized to learn the kernel combination and hashing functions. The obtained hashing functions compress multiple high-dimensional features into tens of binary bits, enabling fast retrieval from a large database. Our approach is extensively validated on 3121 breast-tissue histopathological images by distinguishing between actionable and benign cases. It achieves 88.1% retrieval precision and 91.3% classification accuracy within 16.5 ms query time, comparing favorably with traditional methods.     

SAR image classification based on alpha-stable distribution

We propose a classification algorithm that utilizes the alpha-stable distribution to model the texture features of synthetic aperture radar (SAR) images. The SAR image is first decomposed by stationary wavelet transform (SWT). After that, the alpha-stable distribution is applied to model the high-frequency subband coefficients of the image at each decomposition scale. A regression-type method is then used to estimate the alpha-stable distribution parameters, which form a feature vector that fully describes the texture. Finally, a SAR image classification algorithm is derived by exploiting this feature vector based on the support vector machines (SVM) approach. Because different combinations of alpha-stable distribution parameters contribute to differences in classification precision, a multilevel SVM (MSVM) classification algorithm is also presented to address the issue. Experimental results indicate that the proposed SAR image classification algorithm is effective and the MSVM algorithm improves the classification performance. Moreover, our proposed algorithm has low computational cost as only a small number of the alpha-stable distribution parameters are processed.     

基于遗传算法与支持向量机的基因微阵列分析

背景:微阵列数据的特点是样本含量小,而变量数(基因)多达上万个.此时,传统的统计方法往往因为高维而失效了.遗传算法和支持向量机是近年来发展迅速的机器学习算法,具有很好的分类效果与降维优势.目的:提出将遗传算法与支持向量机结合起来对样本进行分类,并与直接采用支持向量机、筛选差异表达基因后采用支持向量机的结果进行比较.方法:采用Bioconductor提供的数据集golub,它是白血病微阵列芯片实验所得的基因表达数据集,对全部基因采用支持向量机进行分类.采用SAM软件对芯片数据的显著性分析确定不同的差异表达基因并估计错误发现率FDR,以筛选出的76个差异表达基因作为特征基因子集,再采用支持向量机进行分类.将筛选出的76个差异表达基因作为初始的特征基因集合,采用遗传算法一支持向量机再次进行特征基因选择,提高分类准确度,并与全部基因直接采用支持向量机、筛选差异表达基因后采用支持向量机的结果进行比较.同时也对特征基因在代谢通路上的分布和功能作了一定的研究.结果与结论:通过遗传算法降维可以提高支持向量机的分类准确率,特别是剔除了数据中的大量无关基因和噪声,使得经过特征选择后分类准确率提高.结果显示遗传算法与支持向量机结合方法对分类更加有效.此外,通路分析结果显示特征基因的主要功能体现在信号传导和氨基酸代谢上.     

Visualization of nonlinear kernel models in neuroimaging by sensitivity maps

There is significant current interest in decoding mental states from neuroimages. In this context kernel methods, e.g., support vector machines (SVM) are frequently adopted to learn statistical relations between patterns of brain activation and experimental conditions. In this paper we focus on visualization of such nonlinear kernel models. Specifically, we investigate the sensitivity map as a technique for generation of global summary maps of kernel classification models. We illustrate the performance of the sensitivity map on functional magnetic resonance (fMRI) data based on visual stimuli. We show that the performance of linear models is reduced for certain scan labelings/categorizations in this data set, while the nonlinear models provide more flexibility. We show that the sensitivity map can be used to visualize nonlinear versions of kernel logistic regression, the kernel Fisher discriminant, and the SVM, and conclude that the sensitivity map is a versatile and computationally efficient tool for visualization of nonlinear kernel models in neuroimaging.