韵律特征在概率线性判别分析说话人确认中的应用

研究韵律特征在说话人确认中的应用。将整个韵律轨迹以固定段长和段移进行片段划分,并对其进行勒让德多项式拟合从而获取连续性的韵律特征,将特征映射到总变化因子空间,并用概率线性判别分析来补偿说话人和场景的差异。在美国国家标准技术研究院2010年说话人识别评测扩展核心测试集5的基础上加入噪声构造测试集,并分别对韵律特征和传统Mel频率倒谱系数进行测试。结果显示,随着信噪比的逐渐减小,Mel频率倒谱系数性能出现大幅度下降,而韵律特征性能相对比较稳定,两种特征融合后能使系统性能得到进一步提升,等错率和最小检测错误代价相对于Mel频率倒谱系数单系统最多能分别下降9%和11%。实验表明,韵律特征应用于说话人识别中具有较强的噪声鲁棒性,且与传统的Mel频率倒谱系数存在较强的互补性。     

最小方差无失真响应感知倒谱系数在说话人识别中的应用

研究最小方差无失真响应感知倒谱系数在说话人识别中的应用。提取最小方差无失真响应感知倒谱系数,对其进行高斯混合模型建模并采用联合因子分析的方法来拟合高斯混合模型中的说话人和信道差异,在美国国家标准技术研究院2008年说话人识别评测核心测试集上分别对最小方差无失真响应感知倒谱系数和传统的Mel频率倒谱系数进行测试。结果显示,两种不同特征的系统性能相当,采用线性融合方法后,在不同测试集上的等错误率相对下降了7.6%~30.5%,最小检测错误代价相对下降了3.2%~21.2%。实验表明,最小方差无失真响应感知倒谱系数能有效应用于说话人识别中,且与传统的Mel频率倒谱系数存在一定程度的互补性。      

基于仿生特征模式匹配的海洋动物声信号识别*

针对小数据量的海洋动物声信号混合识别,将声信号同态分析过程中的线性频率转换为Mel频率,得到模拟人耳听觉特性的Mel频率倒谱系数作为声信号的特征。按照声信号所属的物种建立特征模板,使用动态时间规整算法对待识别特征进行分类识别,并对特征库和识别算法进行优化。分别提取了6种鱼类、3种虾类、12种鲸类的Mel频率倒谱系数,为每个物种建立特征模板。分3次对3种、5种、6种鱼类进行识别,分别获得了100%、96.25%、94.68%的识别率。对6种鱼类、3种虾类、12种鲸类共21个物种进行混合识别,总识别率由87.56%提升至优化后的88.96%。实验结果表明,基于Mel频率倒谱系数和动态时间规整算法的海洋动物声信号混合识别能够在小数据量时获得较高的识别率,优化后的特征库和识别算法能够提升识别率。     

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

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。     

声门下共鸣的谱规整用于非特定人的语音识别

提出在参数的提取过程中用不同的感知规整因子对不同人的参数归一化,从而实现在非特定人语音识别中对不同人的归一化处理。感知规整因子是基于声门上和声门下之间耦合作用产生声门下共鸣频率来估算的,与采用声道第三共振峰作为基准频率的方法比较,它能较多的滤除语义信息的影响,更好地体现说话人的个性特征。本文提取抗噪性能优于Mel倒谱参数的感知最小方差无失真参数作为识别特征,语音模型用经典的隐马尔可夫模型(HMM)。实验证明,本文方法与传统的语音识别参数和用声道第三共振峰进行谱规整的方法相比,在干净语音中单词错误识别率分别下降了4%和3%,在噪声环境下分别下降了9%和5%,有效地改善了非特定人语音识别系统的性能。      

一种适于说话人识别的非线性频率尺度变换

传统的非线性频率尺度变换虽然能够反映人类听觉系统(HAS:Human Auditory System)的感知特性,但不能区别对待语音中包含的语义和个性特征,在表达说话人个性特征方面并不充分.通过分析语旨信号不同频带短时谱对说话人识别性能的影响,采用最小二乘法多项式曲线拟合技术,提出了一种非线性频率尺度变换.实验表明,与传统的Mel、Bark和ERB频率尺度变换相比,在同样的训练与测试条件下,平均误识率分别降低70.5%,60.8%和70.5%.这一结果说明,本文提出的非线性频率尺度变换有效地增强了短时谱的说话人个性特征,能够提高说话人识别系统的性能.     

基于Transformer编码器的合成语声检测系统

自动说话人认证系统是一种常用的目标说话人身份认证方案,但它在合成语声的攻击下表现出脆弱性,合成语声检测系统试图解决这一问题。该文提出了一种基于Transformer编码器的合成语声检测方法,利用自注意力机制学习输入特征内部的长期依赖关系。合成语声检测问题并不关注句子的抽象语义特征,用参数量较小的模型也能得到较好的检测性能。该文分别测试了4种常用合成语声检测特征在Transformer编码器上的表现,在国际标准的ASVspoof2019挑战赛的逻辑攻击数据集上,基于线性频率倒谱系数特征和Transformer编码器的系统等错误率与串联检测代价函数分别为3.13%和0.0708,且模型参数量仅为0.082 M,在较小参数量下得到了较好的检测性能。     

基于一维卷积神经网络的光纤周界入侵模式识别

针对干涉型分布式光纤传感系统,在通过Mel倒谱系数方法提取扰动信号频域特征进行模式识别的研究基础上,提出了一种基于一维卷积神经网络的光纤入侵模式识别方法。利用还原信号的分级阈值判断并提取入侵信号,有效减少了分帧方法导致的计算时间;构建了基于入侵信号傅里叶变换后的频域信息的一维卷积神经网络,自适应地提取扰动的信号频域特征。搭建了基于直线型Sagnac干涉结构的入侵检测系统,利用大量实验采集的样本数据集对网络进行训练,得到了较好的分类识别结果,测试集的平均识别率达到了96.5%,并对训练后网络的卷积核以及经过卷积核后的入侵信号进行了分析。zscore标准化后,一维卷积神经网络能够识别信号频域中的部分特征,对频率成分复杂的树枝拍打信号识别效果提升较大。     

水下声目标的梅尔倒谱系数智能分类方法

传统水下声目标识别分类方法具有较强的人机交互特性,无法满足未来水下无人平台智能识别分类水声目标的需求。针对这一问题,提出了一种基于梅尔倒谱系数（MFCC）的水下声目标智能识别分类方法,该方法通过提取水下声目标梅尔倒谱系数特征,采用长短时记忆网络（LSTM）构建了智能识别分类模型。使用实际水声信号对该方法进行了验证,结果表明,基于梅尔倒谱系数的水下声目标智能识别分类方法能够在不依赖人工提取特征的情况下,对目标噪声进行识别分类,具备智能化识别分类能力。     

倒谱参数稀疏分解下的汉语音谎言检测

为了提高汉语语音的谎言检测准确率,提出了一种对信号倒谱参数进行稀疏分解的方法。首先,采用小波包滤波器组对语音信号进行多频带划分,求得子频带对数能量并进行离散余弦变换以提取小波包频带倒谱系数,结合梅尔频率谱系数得到倒谱参数;其次,依据K-奇异值分解方法分别利用说谎和非说谎两种状态下的语音倒谱参数集训练得到过完备混合字典,在此字典上根据正交匹配追踪算法对参数集进行稀疏编码提取稀疏特征;最终进行多种分类模型下的识别实验·实验结果表明,稀疏分解方法相比传统参数降维方法具有更好的优化性能,本文推荐的稀疏谱特征最佳识别率达到78.34%,优于其他特征参数,显著提高了谎言检测识别准确率。      

Perceptual MVDR-based cepstral coefficients(PMCCs)for speaker recognition

A feature extraction technique named perceptual MVDR-based cepstral coefficients (PMCCs) was introduced into speaker recognition.PMCCs are extracted and modeled using Gaussian Mixture Models(GMMs) for speaker recognition.In order to compensate for speaker and channel variability effects,joint factor analysis(JFA) is used.The experiments are carried out on the core conditions of NIST 2008 speaker recognition evaluation data.The experimental results show that the systems based on PMCCs can achieve comparable performance to those based on the conventional MFCCs.Besides,the fusion of the two kinds of systems can make significant performance improvement compared to the MFCCs system alone,reducing equal error rate(EER) by the factor between 7.6%and 30.5%as well as minimum detect cost function (minDCF) by the factor between 3.2%and 21.2%on different test sets.The results indicate that PMCCs can be effectively applied in speaker recognition and they are complementary with MFCCs to some extent.     

Fusion of deep shallow features and models for speaker recognition

In order to further improve the performance of speaker recognition, features fusion and models fusion are proposed. The features fusion method is to fuse deep and shallow features. The fused feature describes speaker characteristics more comprehensively than a single feature because of the complementarity between different levels of features. The models fusion method is to fuse i-vectors extracted from different speaker recognition systems. The fused model can combine advantages of different speaker recognition systems. Experimental results show the effectiveness of the proposed methods. Compared with the state-of-the-art system on CASIA North and South dialect corpus,the proposed features fusion system and models fusion system achieved about 54.8% and 69.5% relative improvement on the equal error rate(EER),respectively.     

长时语音特征在说话人识别技术上的应用

本文除介绍常用的说话人识别技术外,主要论述了一种基于长时时频特征的说话人识别方法,对输入的语音首先进行VAD处理,得到干净的语音后,对其提取基本时频特征。在每一语音单元内把基频、共振峰、谐波等时频特征的轨迹用Legendre多项式拟合的方法提取出主要的拟合参数,再利用HLDA的技术进行特征降维,用高斯混合模型的均值超向量表示每句话音时频特征的统计信息。在NIST06说话人1side-1side说话人测试集中,取得了18.7%的等错率,与传统的基于MFCC特征的说话人系统进行融合,等错率从4.9%下降到了4.6%,获得了6%的相对等错率下降。      

Automatic classification and speaker identification of African elephant (Loxodonta africana) vocalizations

A hidden Markov model (HMM) system is presented for automatically classifying African elephant vocalizations. The development of the system is motivated by successful models from human speech analysis and recognition. Classification features include frequency-shifted Mel-frequency cepstral coefficients (MFCCs) and log energy, spectrally motivated features which are commonly used in human speech processing. Experiments, including vocalization type classification and speaker identification, are performed on vocalizations collected from captive elephants in a naturalistic environment. The system classified vocalizations with accuracies of 94.3% and 82.5% for type classification and speaker identification classification experiments, respectively. Classification accuracy, statistical significance tests on the model parameters, and qualitative analysis support the effectiveness and robustness of this approach for vocalization analysis in nonhuman species.     

Analysis of acoustic parameters for consonant voicing classification in clean and telephone speech

This paper describes acoustic cues for classification of consonant voicing in a distinctive feature-based speech recognition system. Initial acoustic cues are selected by studying consonant production mechanisms. Spectral representations, band-limited energies, and correlation values, along with Mel-frequency cepstral coefficients features (MFCCs) are also examined. Analysis of variance is performed to assess relative significance of features. Overall, 82.2%, 80.6%, and 78.4% classification rates are obtained on the TIMIT database for stops, fricatives, and affricates, respectively. Combining acoustic parameters with MFCCs shows performance improvement in all cases. Also, performance in the NTIMIT telephone channel speech shows that acoustic parameters are more robust than MFCCs.     

短时频谱通用背景模型群联合韵律的年龄语音转换

提出一种短时频谱通用背景模型群与韵律参数相结合进行年龄语音转换的方法。谱参数转换方面,同一年龄段各说话者提取语音短时谱系数并建立高斯混合模型,然后依据语音特征相似性对说话者进行聚类,每一类训练一个通用背景模型,最终得到通用背景模型群和一组短时频谱转换函数。谱参数转换之后再对共振峰进一步微调。韵律参数转换方面,基频和语速分别建立单高斯和平均时长率模型来推导转换函数。实验结果显示,提出的方法在ABX和MOS等评价指标上比传统的双线性法有明显的优势,相对单一通用背景模型法的对数似然度变化率提高了4%。这一结果表明提出的方法能够使转换语音具有良好目标倾向性的同时有较好的语音质量,性能较传统方法有明显提升。      

Spectro-temporal modulation subspace-spanning filter bank features for robust automatic speech recognition

In an attempt to increase the robustness of automatic speech recognition (ASR) systems, a feature extraction scheme is proposed that takes spectro-temporal modulation frequencies (MF) into account. This physiologically inspired approach uses a two-dimensional filter bank based on Gabor filters, which limits the redundant information between feature components, and also results in physically interpretable features. Robustness against extrinsic variation (different types of additive noise) and intrinsic variability (arising from changes in speaking rate, effort, and style) is quantified in a series of recognition experiments. The results are compared to reference ASR systems using Mel-frequency cepstral coefficients (MFCCs), MFCCs with cepstral mean subtraction (CMS) and RASTA-PLP features, respectively. Gabor features are shown to be more robust against extrinsic variation than the baseline systems without CMS, with relative improvements of 28% and 16% for two training conditions (using only clean training samples or a mixture of noisy and clean utterances, respectively). When used in a state-of-the-art system, improvements of 14% are observed when spectro-temporal features are concatenated with MFCCs, indicating the complementarity of those feature types. An analysis of the importance of specific MF shows that temporal MF up to 25 Hz and spectral MF up to 0.25 cycles/channel are beneficial for ASR.     

深浅层特征及模型融合的说话人识别

为了进一步提高说话人识别系统的性能,提出基于深、浅层特征融合及基于I-Vector的模型融合的说话人识别。基于深、浅层特征融合的方法充分考虑不同层级特征之间的互补性,通过深、浅层特征的融合,更加全面地描述说话人信息;基于I-Vector模型融合的方法融合不同说话人识别系统提取的I-Vector特征后进行距离计算,在系统的整体结构上综合了不同说话人识别系统的优势。通过利用CASIA南北方言语料库进行测试,以等错误率为衡量指标,相比基线系统,基于深、浅层特征融合的说话人识别其等错误率相对下降了54.8%,基于I-Vector的模型融合的方法其等错误率相对下降了69.5%。实验结果表明,深、浅层特征及模型融合的方法是有效的。      

Classification of stop place in consonant-vowel contexts using feature extrapolation of acoustic-phonetic features in telephone speech

Knowledge-based speech recognition systems extract acoustic cues from the signal to identify speech characteristics. For channel-deteriorated telephone speech, acoustic cues, especially those for stop consonant place, are expected to be degraded or absent. To investigate the use of knowledge-based methods in degraded environments, feature extrapolation of acoustic-phonetic features based on Gaussian mixture models is examined. This process is applied to a stop place detection module that uses burst release and vowel onset cues for consonant-vowel tokens of English. Results show that classification performance is enhanced in telephone channel-degraded speech, with extrapolated acoustic-phonetic features reaching or exceeding performance using estimated Mel-frequency cepstral coefficients (MFCCs). Results also show acoustic-phonetic features may be combined with MFCCs for best performance, suggesting these features provide information complementary to MFCCs.