基于语音频谱融合特征的手机来源识别
|
摘要
随着手机录音设备的普及以及各种功能强大且易于操作的数字媒体编辑软件的出现,语音的手机来源识别已成为多媒体取证领域重要的热点问题,针对该问题提出了一种基于频谱融合特征的手机来源识别算法。首先,通过分析不同手机相同语音的语谱图,发现不同手机的语音频谱特征是不同的;然后对语音的频谱信息量、对数谱和相位谱特征进行了研究;其次,将三个特征串联构成原始融合特征,并用每个样本的原始融合特征构建样本特征空间;最后,采用WEKA平台的CfsSubsetEval评价函数按照最佳优先搜索原则对所构建的特征空间进行特征选择,并采用LibSVM对特征选择后的样本特征空间进行模型训练和样本识别。实验部分给出了特征选择后的频谱单一特征和频谱融合特征在23款主流型号的手机语音库上分类的结果。实验结果表明,该算法使用频谱融合特征有效提高了手机品牌类内的平均识别准确率,在TIMIT翻录语音数据库和自建的CKC-SD语音数据库上分别达到99.96%和99.91%;另外,与Hanilci基于梅尔倒谱系数特征的录音设备来源识别算法进行了对比,平均识别准确率分别提高了6.58和5.14个百分点。因此可得本文所提特征可有效提高平均识别准确率,降低手机类内识别的误判率。
With the popularity of cell-phone recording devices and the availability of various powerful and easy to operate digital media editing software, source cell-phone identification has become a hot topic in multimedia forensics, a cell-phone source recognition algorithm based on spectral fusion features was proposed to solve this problem. Firstly, the same speech spectrograms of different cell-phones were analyzed, it was found that the speech spectral characteristics of different cellphones were different; then the logarithmic spectrum, phase spectrum and information quantity for a speech were researched.Secondly, the three features were connected in series to form the original fusion feature, and the sample feature space was constructed with the original fusion feature of each sample. Finally, the evaluation function CfsSubsetEval of WEKA platform was selected according to the best priority search method to select features, and LibSVMwas used to model training and sample recognition after feature selection. Twenty-three popular cell-phone models were evaluated in the experiment, the results showed that the proposed spectral fusion feature has higher identification accuracy for cell-phone brands than spectral single feature and the average identification accuracies achieved 99. 96% and 99. 91% on TIMIT database and CKC-SD database. In addition, it was compared with the source identification algorithm of Hanilci based on Mel frequency cepstral coefficients, the average identification accuracy was improved by 6. 58 and 5. 14 percentage points respectively. Therefore, the proposed algorithm can improve the average identification accuracy and effectively reduce the false positives rate of cell-phone source identification.
With the popularity of cell-phone recording devices and the availability of various powerful and easy to operate digital media editing software, source cell-phone identification has become a hot topic in multimedia forensics, a cell-phone source recognition algorithm based on spectral fusion features was proposed to solve this problem. Firstly, the same speech spectrograms of different cell-phones were analyzed, it was found that the speech spectral characteristics of different cellphones were different; then the logarithmic spectrum, phase spectrum and information quantity for a speech were researched.Secondly, the three features were connected in series to form the original fusion feature, and the sample feature space was constructed with the original fusion feature of each sample. Finally, the evaluation function CfsSubsetEval of WEKA platform was selected according to the best priority search method to select features, and LibSVMwas used to model training and sample recognition after feature selection. Twenty-three popular cell-phone models were evaluated in the experiment, the results showed that the proposed spectral fusion feature has higher identification accuracy for cell-phone brands than spectral single feature and the average identification accuracies achieved 99. 96% and 99. 91% on TIMIT database and CKC-SD database. In addition, it was compared with the source identification algorithm of Hanilci based on Mel frequency cepstral coefficients, the average identification accuracy was improved by 6. 58 and 5. 14 percentage points respectively. Therefore, the proposed algorithm can improve the average identification accuracy and effectively reduce the false positives rate of cell-phone source identification.
引文
[1]胡永健,刘琲贝,贺前华.数字多媒体取证技术综述[J].计算机应用,2010,30(3):657-662.(HU Y J,LIU B B,HE Q H.Survey on techniques of digital multimedia forensics[J].Journal of Computer Applications,2010,30(3):657-662.)
[2]ESKIDERE O.Identifying acquisition devices from recorded speech signals using wavelet based features[J].Turkish Journal of Electrical Engineering&Computer Sciences,2015,24:1942-1954.
[3]贺前华,王志锋,RUDNICKY A I,等.基于改进PNCC特征和两步区分性训练的录音设备识别方法[J].电子学报,2014,42(1):191-198.(HE Q H,WANG Z F,RUDNICKY A I,et al.Arecording device identification algorithm based on improved PNCCfeature and two-step discriminative training[J].Acta Electronica Sinica,2014,42(1):191-198.)
[4]KOTROPOULOS C,SAMARAS S.Mobile phone identification using recorded speech signals[C]//Proceedings of the 2014 19th International Conference on Digital Signal Processing.Piscataway,NJ:IEEE,2014:586-591.
[5]ESKIDERE O.Source microphone identification from speech recordings based on a Gaussian mixture model[J].Turkish Journal of Electrical Engineering&Computer Sciences,2014,22(3):754-767.
[6]PANAGAKIS Y,KOTROPOULOS C L.Telephone handset identification by collaborative representations[J].International Journal of Digital Crime&Forensics,2013,5(4):1-14.
[7]HICSONMEZ S,SENCAR H T,AVCIBAS I.Audio codec identification from coded and transcoded audios[J].Digital Signal Processing,2013,23(5):1720-1730.
[8]裴安山,王让定,严迪群.基于设备本底噪声频谱特征的手机来源识别[J].电信科学,2017,33(1):85-94.(PEI A S,WANG R D,YAN D Q.Cell-phone origin identification based on spectral features of device self-noise[J].Telecommunications Science,2017,33(1):85-94.)
[9]裴安山,王让定,严迪群.基于语音静音段特征的手机来源识别方法[J].电信科学,2017,33(7):103-111.(PEI A S,WANG R D,YAN D Q.Source cell-phone identification from recorded speech using non-speech segments[J].Telecommunications Science,2017,33(7):103-111.)
[10]HANILCI C,ERTAS F,ERTAS T,et al.Recognition of brand and models of cell-phones from recorded speech signals[J].IEEETransactions on Information Forensics&Security,2012,7(2):625-634.
[11]KOTROPOULOS C L.Source phone identification using sketches of features[J].IET Biometrics,2014,3(2):75-83.
[12]沈连丰,叶之慧.信息论与编码[M].北京:科学出版社.2004:12-17.(SHEN L F,YE Z H.Information Theory and Coding[M].Beijing:Science Press,2004:12-17.)
[13]XU L,YAN P,CHANG T.Best first strategy for feature selection[C]//Proceedings of the 9th International Conference on Pattern Recognition.Piscataway,NJ:IEEE,1988:706-708.
[14]HALL M A.Correlation-based feature selection for machine learning[D].Hamilton,New Zealand:The University of Waikato,1999:51-74.
[15]林升梁,刘志.基于RBF核函数的支持向量机参数选择[J].浙江工业大学学报,2007,35(2):163-167.(LIN S L,LIU Z.Parameter selection in SVM with RBF kernel function[J].Journal of Zhejiang University of Technology,2007,35(2):163-167.)
[2]ESKIDERE O.Identifying acquisition devices from recorded speech signals using wavelet based features[J].Turkish Journal of Electrical Engineering&Computer Sciences,2015,24:1942-1954.
[3]贺前华,王志锋,RUDNICKY A I,等.基于改进PNCC特征和两步区分性训练的录音设备识别方法[J].电子学报,2014,42(1):191-198.(HE Q H,WANG Z F,RUDNICKY A I,et al.Arecording device identification algorithm based on improved PNCCfeature and two-step discriminative training[J].Acta Electronica Sinica,2014,42(1):191-198.)
[4]KOTROPOULOS C,SAMARAS S.Mobile phone identification using recorded speech signals[C]//Proceedings of the 2014 19th International Conference on Digital Signal Processing.Piscataway,NJ:IEEE,2014:586-591.
[5]ESKIDERE O.Source microphone identification from speech recordings based on a Gaussian mixture model[J].Turkish Journal of Electrical Engineering&Computer Sciences,2014,22(3):754-767.
[6]PANAGAKIS Y,KOTROPOULOS C L.Telephone handset identification by collaborative representations[J].International Journal of Digital Crime&Forensics,2013,5(4):1-14.
[7]HICSONMEZ S,SENCAR H T,AVCIBAS I.Audio codec identification from coded and transcoded audios[J].Digital Signal Processing,2013,23(5):1720-1730.
[8]裴安山,王让定,严迪群.基于设备本底噪声频谱特征的手机来源识别[J].电信科学,2017,33(1):85-94.(PEI A S,WANG R D,YAN D Q.Cell-phone origin identification based on spectral features of device self-noise[J].Telecommunications Science,2017,33(1):85-94.)
[9]裴安山,王让定,严迪群.基于语音静音段特征的手机来源识别方法[J].电信科学,2017,33(7):103-111.(PEI A S,WANG R D,YAN D Q.Source cell-phone identification from recorded speech using non-speech segments[J].Telecommunications Science,2017,33(7):103-111.)
[10]HANILCI C,ERTAS F,ERTAS T,et al.Recognition of brand and models of cell-phones from recorded speech signals[J].IEEETransactions on Information Forensics&Security,2012,7(2):625-634.
[11]KOTROPOULOS C L.Source phone identification using sketches of features[J].IET Biometrics,2014,3(2):75-83.
[12]沈连丰,叶之慧.信息论与编码[M].北京:科学出版社.2004:12-17.(SHEN L F,YE Z H.Information Theory and Coding[M].Beijing:Science Press,2004:12-17.)
[13]XU L,YAN P,CHANG T.Best first strategy for feature selection[C]//Proceedings of the 9th International Conference on Pattern Recognition.Piscataway,NJ:IEEE,1988:706-708.
[14]HALL M A.Correlation-based feature selection for machine learning[D].Hamilton,New Zealand:The University of Waikato,1999:51-74.
[15]林升梁,刘志.基于RBF核函数的支持向量机参数选择[J].浙江工业大学学报,2007,35(2):163-167.(LIN S L,LIU Z.Parameter selection in SVM with RBF kernel function[J].Journal of Zhejiang University of Technology,2007,35(2):163-167.)