共查询到18条相似文献,搜索用时 78 毫秒
1.
在光谱分辨率增强技术中,线型优化最大熵谱估计方法LOMEE(line shape optimized maximum entropy spectral estimation)是指通过傅里叶自退卷积技术消除光谱谱线线型对干涉图的影响,而后对干涉图进行自回归参数建模,求出自回归模型系数,代入光谱估计公式得到光谱图。文章采用了两种方法:修正协方差法MCOV(modified covariance method))和伯格法(Burg method)求解线型优化最大熵谱估计方法中的自回归模型系数,通过仿真试验,将阶次、信噪比对估计光谱的影响做了详尽的比较。研究结果表明,在线型优化最大熵谱估计方法中,用修正协方差法求解自回归模型系数要优于伯格法。 相似文献
2.
光谱分辨率增强方法品质因子研究 总被引:1,自引:2,他引:1
本文提出了一种新的评估光谱复原结果好坏的品质因子 (quality factor).该品质因子综合考虑到了前向线性预测和后向线性预测.研究了在不同噪音情况下,当采用伯格法(Burg Method)来求解自回归模型系数时,品质因子随模型阶次的变化以及对复原光谱的影响,并且与芬兰学者Kauppinen提出的品质因子在同等条件下作了详尽的比较.研究结果表明,本文提出的品质因子更具有优越性. 相似文献
3.
《中国光学与应用光学文摘》2007,(5)
O433 2007054095线型优化最大熵线性预测方法自回归模型三种求解方法的比较=Comparison of three methods to solve AR parameters of line shape opti mized maxi mumentropy linear prediction[刊,中]/高晓峰(中科院西安光机所.陕西,西安(710068)) ,相里斌//光子学报.? 2007 相似文献
4.
假设信号和噪声是独立的平稳彩色高斯过程,本文用独立脉冲对噪声抵消法的自回归谱矩估计值(PPARNC)和协方差函数谱矩估计值(PPCNC)以及独立脉冲对最大似然法的自回归谱矩估计值(PPARML)和协方差函数谱矩估计值(PPCML)分析了信号的频率估计值和带宽估计值、即一阶谱矩和二阶谱矩估计值的均值和方差。分析表明,PPARNC和PPCNC均优于PPARML和PPCML,其中PPARNC明显优于其它方法。 相似文献
5.
基于最大似然多项式回归的鲁棒语音识别 总被引:2,自引:0,他引:2
本文针对最大似然线性回归算法线性假设的缺点,将多项式回归方法用于模型自适应,构建了基于最大似然多项式回归的非线性模型自适应算法。该算法在对数谱域用多项式回归方法,逼近每个Mel子带上识别环境模型均值与训练环境模型均值之间的非线性关系。多项式系数通过EM算法和最大似然准则从识别环境下的少量自适应数据中估计。实验结果表明,二阶多项式就可以较好地逼近模型均值的非线性环境变换关系。在噪声补偿和说话人自适应实验中,最大似然多项式回归算法的误识率都明显低于最大似然线性回归算法。本文算法较好地克服了线性模型自适应算法线性假设的缺陷,可同时减小噪声,和说话人的改变或其它因素对语音识别系统的影响,尤其适合说话人和噪声的联合自适应。 相似文献
6.
7.
文本无关说话人识别的全特征矢量集模型及互信息评估方法 总被引:1,自引:0,他引:1
提出了一种文本无关说话人识别的全特征矢量集模型及互信息评估方法,该模型通过对一组说话人语音数据在特征空间进行聚类而形成,全面地反映了说话人语音的个性特征。对于说话人语音的似然度计算与判决,则提出了一种互信息评估方法,该算法综合分析距离空间和信息空间的似然度,并运用最大互信息判决准则进行识别判决。实验分析了线性预测倒谱系数(LPCC)和Mel频率倒谱系数(MFCC)两种情况下应用全特征矢量集模型和互信息评估算法的说话人识别性能,并与高斯混合模型进行了比较。结果表明:全特征矢量集模型和互信息评估算法能够充分反映说话人语音特征,并能够有效评估说话人语音特征相似程度,具有很好的识别性能,是有效的。 相似文献
8.
NDIR多组分气体分析的干扰修正方法研究 总被引:1,自引:0,他引:1
介绍了非分散红外(Non-Dispersive Infrared,NDIR)多组分气体分析的一种干扰修正方法.根据逐次积分气体吸收模型和方法,并考虑温度和气压对积分线强和线型的影响,选择洛伦兹展宽线型,得到了CO,CO2,NO和H2O在各个滤波通道的响应系数.利用响应系数建立了用于干扰修正的四元线性回归方程组,通过求解... 相似文献
9.
10.
基于回归分析的语音识别快速自适应算法 总被引:2,自引:2,他引:0
从回归分析的角度推导出最大似然线性回归算法的等价算法--最小二乘线性回归算法,以及相应的多元线性回归模型。该模型中回归因子间存在着多重共线性,它导致了算法在自适应数据很少时失效。为减轻多重共线性的影响,提出改进算法:伪自适应数据算法。实验表明,当仅有1s-3s自适应数据时,新算法使得系统误识率相对下降2%-6%,随着自适应数据增多,其性能与最大似然线性回归(或最小二乘线性同归)算法趋于一致。 相似文献
11.
12.
Modern power spectral estimation technique was applied to the signal analysis of inductively coupled plasma atomic emission spectrometry (ICP-AES) in the present work. Two methods, Levinson-Durbin and Burg methods, were used to calculate auto-regression (AR) model parameters and the comparison was made. the influence of different AR model orders was studied to find the optimal order by calculating AR model parameters and power spectral density. Results showed that the Burg method gave higher resolution. Characteristics of real ICP-AES measurements were studied and compared by the two methods. 相似文献
13.
14.
Modulation of electroencephalograph activity by manual acupuncture stimulation in healthy subjects:An autoregressive spectral analysis 下载免费PDF全文
To investigate whether and how manual acupuncture(MA) modulates brain activities,we design an experiment where acupuncture at acupoint ST36 of the right leg is used to obtain electroencephalograph(EEG) signals in healthy subjects.We adopt the autoregressive(AR) Burg method to estimate the power spectrum of EEG signals and analyze the relative powers in delta(0 Hz-4 Hz),theta(4 Hz-8 Hz),alpha(8 Hz-13 Hz),and beta(13 Hz-30 Hz) bands.Our results show that MA at ST36 can significantly increase the EEG slow wave relative power(delta band) and reduce the fast wave relative powers(alpha and beta bands),while there are no statistical differences in theta band relative power between different acupuncture states.In order to quantify the ratio of slow to fast wave EEG activity,we compute the power ratio index.It is found that the MA can significantly increase the power ratio index,especially in frontal and central lobes.All the results highlight the modulation of brain activities with MA and may provide potential help for the clinical use of acupuncture.The proposed quantitative method of acupuncture signals may be further used to make MA more standardized. 相似文献
15.
Modulation of electroencephalograph activities by manual acupuncture stimulation in healthy subjects: An autoregressive spectral analysis 下载免费PDF全文
To investigate whether and how manual acupuncture (MA) modulates brain activities, we design an experiment that acupuncture at acupoint ST36 of right leg to obtain electroencephalograph (EEG) signals in healthy subjects. We adopt autoregressive (AR) Burg method to estimate the power spectrum of EEG signals and analyze the relative powers in delta (0 Hz-4 Hz), theta (4 Hz-8 Hz), alpha (8 Hz-13 Hz), and beta (13 Hz-30 Hz) bands. Our results show that MA at ST36 can significantly increase the EEG slow wave relative power (delta band) and reduce the fast wave relative powers (alpha and beta bands), while there are no statistical differences in theta band relative power between different acupuncture states. In order to quantify the ratio of slow to fast wave EEG activity, we compute the power ratio index. It is found that the MA can significantly increase the power ratio index, especially in frontal and central lobes. All the results highlight the modulation of brain activities with MA and may provide potential helps in clinical treatment of acupuncture. The proposed quantitative method of acupuncture signals may be further used to make MA more standardized. 相似文献
16.
17.
I. M. Dozmorov 《Russian Physics Journal》1968,11(3):52-53
The Bondi/Van der Burg/Metzner solution of the Einstein field equations for an arbitrary axisymmetric system is examined. Expressions for the energy and momentum of the system are obtained by tetrad formalism. 相似文献