超声多普勒音频信号分形特征研究

本文报道了心动周期内超声多普勒音频信号的分维值计算方法和对超声多普勒信号的分维特征进行统计分析的结果。通过对１００多例胎儿脐血流信号的分析研究,结果表明：分维值对某些病症具有敏感性,甚至优于现有的声谱参数法。     

基于遗传算法和最小二乘支持向量机的胎儿状态智能评估

胎心宫缩图是一种临床常用的评估胎儿健康状况的电子监护技术,具有易受主观因素影响导致诊断率较低的缺点。为降低误诊率,辅助医生做出准确的医疗决策,本文提出了一种基于胎心率信号分析胎儿状态的智能评估方法。首先,本文将来自捷克技术大学—布尔诺大学医院公开数据库的信号进行预处理后,对其中的胎心率信号进行多模态特征提取,然后利用设计的基于k—最近邻遗传算法选择最优特征子集,最后采用最小二乘支持向量机法对其分类。实验结果显示,利用本文提出的方法对胎儿状态进行分类,其准确度可达91%,灵敏度为89%,特异度为94%,质量指标为92%,受试者工作特征曲线下面积为92%,具有较好的分类性能,可辅助临床医生对胎儿状态做出有效评估。     

多变量自回归和多线性主成分分析结合的多通道信号特征提取研究

脑机接口(BCI)系统通过从脑信号中提取特征对其进行识别。针对自回归模型特征提取方法和传统主成分分析降维方法处理多通道信号的局限性,本文提出了多变量自回归(MVAR)模型和多线性主成分分析(MPCA)结合的多通道特征提取方法,并用于脑磁图/脑电图(MEG/EEG)信号识别。首先计算MEG/EEG信号的MVAR模型的系数矩阵,然后采用MPCA对系数矩阵进行降维,最后使用线性判别分析分类器对脑信号分类。创新在于将传统单通道特征提取方法扩展到多通道。选用BCI竞赛IV数据集3和1数据进行实验验证,两组实验结果表明MVAR和MPCA结合的特征提取方法处理多通道信号是可行的。     

节律自适应的运动想象脑电空域特征提取方法

目的：针对运动想象脑电（MI-EEG）信号个体差异性大，特征质量依赖频带的选择，导致多类MI-EEG信号识别效果差的问题，提出节律自适应的空域特征提取方法。方法：用滤波器组共空间模式（FBCSP）提取多个频带的空域特征，结合免疫粒子群优化算法，对特征提取过程中的频、空参数寻优，实现节律、空域特征提取参数的自适应调整，获取最优节律下的FBCSP空域特征，提升多类MI-EEG信号的识别准确率。结果：本文方法在BCI-Ⅳ Dataset 2a、BCI-ⅢDataset 3a数据集上取得85.49%的平均准确率，较原始FBCSP方法提升10.84%。结论：本文方法更好地获取了脑电空域特征，能有效提高分类正确率，为MI-EEG分类提供了新的解决思路。     

基于回声状态网络的脑电信号特征提取

在脑电(EEG)信号自动检测和分类的研究中,EEG信号的特征提取至关重要。本文分析了目前主要EEG信号特征提取方法的优缺点,并提出了一种基于回声状态网络(ESN)的EEG信号特征提取方法。该方法可以实现EEG信号的非线性特征提取,并且其特征提取过程是近似可逆的,因而在特征提取过程中损失的信息较少。该方法在EEG信号特征提取过程中,主要计算量是求解状态矩阵的伪逆,计算简单高效。在对波恩大学癫痫研究所的EEG数据库进行多类别分类的实验中,本文所提出的EEG信号特征提取方法展现出了良好的性能。     

癫痫脑电特征波的综合检测分类方法研究

本文将小波变换、人工神经网络、专家规则判据等多种检测方法有机地结合起来 ,用于癫痫脑电特征波的检测与分类 ,以充分发挥不同方法的优势。这种综合检测分类方法是先将预处理的多导脑电时间序列经小波变换将脑电中癫痫特征波在不同尺度下分离出来 ,再对选出的癫痫嫌疑波进行特征参数提取 ,然后把特征参数送入已经训练好的人工神经网络进行分类识别 ,最后再由专家规则判断筛选并作出检测分类统计报告。研究表明 ,该方法具有很好的信号特征提取和屏蔽随机噪声能力 ,获得了较好的检出率 ;尤其适合于非平稳、非线性生物医学信号的检测分类 ,值得进一步深入研究     

基于特征向量法和支持向量机的抑郁症脑电信号分类

目的:从抑郁症患者EEG信号中提取与疾病相关的信息以实现对抑郁症患者与健康人的自动分类.方法:用特征向量法对抑郁症患者与健康人脑电进行特征提取,得到脑电信号功率谱幅度的最大值、最小值、平均值和标准偏差等特征参数,然后用支持向量机分类器进行训练和分类,并进行测试验证.结果:相对于用小波变换提取的频率相关参数作为分类特征,采用本文特征向量法功率谱估计提取的特征参数为分类特征的分类器具有更好的分类效果,其抑郁症患者和健康人脑电信号的分类准确率可以达到95.6%.结论:该研究成果为抑郁症疾病的物理诊断提供了一种新的途径.     

基于时频分析的个性化脑-机接口设计

为了提高脑—机接口的分类正确率和实时性,本研究提出基于时频分析的个性化脑—机接口设计方案。通过对多导联EEG数据进行离线谱图分析,量化不同导联EEG信号在想象左右手运动时,其各个频带能量随时间的变化规律。然后将各导联EEG信号不同频带能量变化特征在左右手运动想象时的差异进行脑地形图成像。进而根据特征差异在大脑皮层的分布为不同用户的最佳电极位置选择和特征频带的确定提供直接依据。实验结果证明,该方法有效减少了特征提取的盲目性,用最少的导联,最少的特征向量,较简单的特征提取和分类算法,取得很好的分类效果,具有较高实用价值。     

主成分分析和线性判别分析应用于心电信号特征提取和诊断算法研究

针对心脏疾病发病率高且不易自主检测的问题,提出了一种心电信号特征提取和分类诊断算法。首先对心电信号进行提升小波变换和改进半软阈值相结合的预处理变换,在去除心电信号的噪声后,利用主成分分析(principal component analysis,PCA)对心电信号进行降维,并利用核独立成分提取心电信号的非线性特征;同时离散小波变换提取去噪后心电信号的频域特征,基于线性判别分析(linear discriminant analysis, LDA)对频域统计特征进行降维处理。将两种不同的特征向量组成多域特征空间,最后利用支持向量机对多域特征空间分类,遗传算法对其参数进行寻优,从而实现心电信号特征的分类。实验结果表明,所提出的算法能够对5类心电节拍进行准确分类,分类效率达99.11%。     

基于多通道经验模式分解的脑机接口特征提取

针对脑机接口(BCI)系统中的多通道非平稳脑电(EEG)信号和脑磁(MEG)信号,本文提出一种基于多通道经验模式分解(MEMD)与功率特征结合的信号特征提取算法。首先将多通道脑信号经MEMD算法分解为一系列多尺度多元固有模态函数(IMF)近似平稳分量,然后对每个IMF分量提取功率特征,并利用主成分分析(PCA)降维处理,最后使用线性判别分析分类器对信号特征分类。实验采用第三次和第四次国际BCI竞赛的数据进行验证,对皮层EEG信号和MEG信号运动想象任务的识别正确率分别达到92.0%和46.2%,均位于竞赛第一名水平。实验结果表明本文所提方法有较好有效性和稳定性,为脑信号特征提取提供了新思路。     

An automated feature extraction and emboli detection system based on the PCA and fuzzy sets

The Doppler ultrasound technique is commonly used to detect emboli in the cerebral circulation. Here an automated feature extraction and emboli detection system is proposed based on the principal components analysis (PCA) and fuzzy sets. In the system, two features, R(ry) and k, are extracted by the PCA method. Meanwhile, MMR and sigma(f min) are obtained with the traditional temporal processing and spectrogram analysis, respectively. Normal blood flow signals are firstly distinguished from abnormal signals by MMR. Then signals containing emboli and disturbance noises are further differentiated by other features based on fuzzy sets. From experiments with computer-simulated and clinical Doppler ultrasound signals, it is shown that features extracted from the PCA method achieve better classification performance than those of traditional methods. The fuzzy-based detection system not only obtains high classification accuracy but is more applicable in clinical diagnosis.     

超声多普勒血流和管壁搏动信号的分离研究

目的针对超声多普勒血流检测中,传统的高通滤波法在滤除管壁搏动信号的同时也会滤除低频血流信号的问题,本研究提出一种以心电信号（electrocardiography,ECG）作为参考信号的自适应滤波的方法消除管壁干扰。方法包括两方面：其一,采用心电信号作为参考信号对超声多普勒信号进行自适应滤波;其二,采用多级自适应滤波并选择不同的参考信号的滤波方案。分别使用上述方法和高通滤波法对仿真的超声多普勒信号进行处理,并将结果进行比较。结果与传统的高通滤波法相比,该方法在有效抑制管壁搏动信号的同时保留一部分低频血流信号成分。结论该方法能较准确地提取出完整的血流超声多普勒信号,具有一定的临床应用价值。     

Comparison of eigenvector methods with classical and model-based methods in analysis of internal carotid arterial Doppler signals

Doppler ultrasound is known as a reliable technique, which demonstrates the flow characteristics and resistance of arteries in various vascular disease. In this study, internal carotid arterial Doppler signals recorded from 105 subjects were processed by PC-computer using classical, model-based, and eigenvector methods. The classical method (fast Fourier transform), two model-based methods (Burg autoregressive, least-squares modified Yule-Walker autoregressive moving average methods), and three eigenvector methods (Pisarenko, multiple signal classification, and Minimum-Norm methods) were selected for processing internal carotid arterial Doppler signals. Doppler power spectra of internal carotid arterial Doppler signals were obtained using these spectrum analysis techniques. The variations in the shape of the Doppler power spectra were examined in order to obtain medical information. These power spectra were then used to compare the applied methods in terms of their frequency resolution and the effects in determination of stenosis and occlusion in internal carotid arteries.     

Vectorial reconstruction of retinal blood flow in three dimensions measured with high resolution resonant Doppler Fourier domain optical coherence tomography

Resonant Doppler Fourier domain optical coherence tomography (FDOCT) is a functional imaging tool for extracting tissue flow. The method is based on the effect of interference fringe blurring in spectrometer-based FDOCT, where the path difference between structure and reference changes during camera integration. If the reference path length is changed in resonance with the Doppler frequency of the sample flow, the signals of resting structures will be suppressed, whereas the signals of blood flow are enhanced. This allows for an easy extraction of vascularization structure. Conventional flow velocity analysis extracts only the axial flow component, which strongly depends on the orientation of the vessel with respect to the incident light. We introduce an algorithm to extract the vessel geometry within the 3-D data volume. The algorithm calculates the angular correction according to the local gradients of the vessel orientations. We apply the algorithm on a measured 3-D resonant Doppler dataset. For validation of the reproducibility, we compare two independently obtained 3-D flow maps of the same volunteer and region.     

超声多普勒血流信号的分析方法

超声多普勒技术是无损诊断血管疾病的一种重要手段,因此对超声多普勒血流信号的分析处理可以为疾病诊断提供重要依据.为了分析和处理像超声多普勒这类非平稳信号,人们对基于傅立叶变换的传统信号分析方法进行了推广乃至根本性的革命,提出并发展了一系列新的信号分析理论.本文对应用于超声多普勒血流信号分析的短时傅立叶变换、小波变换、参数模型法和Cohen类的时频分布等方法作了着重论述.     

Doppler ultrasound wall removal based on the spatial correlation of wavelet coefficients

In medical Doppler ultrasound systems, a high-pass filter is commonly used to reject echoes from the vessel wall. However, this leads to the loss of the information from the low velocity blood flow. Here a spatially selective noise filtration algorithm cooperating with a threshold denoising based on wavelets coefficients is applied to estimate the wall clutter. Then the blood flow signal is extracted by subtracting the wall clutter from the mixed signal. Experiments on computer simulated signals with various clutter-to-blood power ratios indicate that this method achieves a lower mean relative error of spectrum than the high-pass filtering and other two previously published separation methods based on the recursive principle component analysis and the irregular sampling and iterative reconstruction, respectively. The method also performs well when applied to in vivo carotid signals. All results suggest that this approach can be implemented as a clutter rejection filter in Doppler ultrasound instruments.     

Versatile microcomputer-based system for the capture,storage and processing of spectrum-analysed Doppler ultrasound blood flow signals

The evaluation of any method of analysis of Doppler ultrasound blood flow signals is involved and time consuming because of the considerable time necessary to investigate a statistically significant representative population of arteriopathic blood flow waveforms. To overcome these problems we have developed a microcomputer-based system for the capture, storage and processing of spectrum-analysed Doppler ultrasound blood flow signals. This system allows the collection and storage on floppy disk of waveforms from many sites in a large population of arteriopaths and their later analysis using any desired method. Having thus created on disk a suitable population of arteriopathic waveforms the evaluation of any method of waveform analysis, whether existing or new, is a much more convenient and far less time-consuming process. The system described is extremely versatile, for example in addition to the collection of data for postprocessing the system is also used for the real-time analysis of blood flow waveforms.     

Correction for broadening in Doppler blood flow spectrum estimated using wavelet transform

The conventionally used spectral estimation technique for Doppler blood flow signal analysis is short-time Fourier transform (STFT). But this method requires stationarity of the signal during the window interval. Wavelet transform (WT), which has a flexible time-frequency window, is particularly suitable for nonstationary signals. In recently years, the WT has been used to investigate its advantages and limitations for the analysis of Doppler blood flow signals. In these studies, the estimated spectral width of Doppler blood flow signals using the WT might include significant window and nonstationarity broadening errors. These broadening errors of the time-varying spectrum were clearly undesirable since it would tend to mask the effect of flow disturbance on the spectra width. In this paper, a closed form expression for window and nonstationary root-mean-squared (rms) spectral width is given when using the WT to estimate the Doppler blood flow spectrum. The increases in the rms spectral width can be calculated and then the spectral width estimation based on the WT can be corrected.     

Noise reduction in Doppler ultrasound signals using an adaptive decomposition algorithm

A novel de-noising method for improving the signal-to-noise ratio (SNR) of Doppler ultrasound blood flow signals, called the matching pursuit method, has been proposed. Using this method, the Doppler ultrasound signal was first decomposed into a linear expansion of waveforms, called time-frequency atoms, which were selected from a redundant dictionary named Gabor functions. Subsequently, a decay parameter-based algorithm was employed to determine the decomposition times. Finally, the de-noised Doppler signal was reconstructed using the selected components. The SNR improvements, the amount of the lost component in the original signal and the maximum frequency estimation precision with simulated Doppler blood flow signals, have been used to evaluate a performance comparison, based on the wavelet, the wavelet packets and the matching pursuit de-noising algorithms. From the simulation and clinical experiment results, it was concluded that the performance of the matching pursuit approach was better than those of the DWT and the WPs methods for the Doppler ultrasound signal de-noising.     

用超声方法检测血栓的物理实验研究

本文使用基于ＴＭＳ３２０Ｃ２５高速信号处理卡的双通道脉冲破经颅多普勒血流分析仪,建立起模拟血栓实验平台,从频域和时域两方面采集模拟血栓数据,并对血栓的特性及其检测方法进行研究,实验结果表明,从时域延迟和频谱特征变化两方面来检测血栓是一种应用前景的方法。