《计算机应用研究》|Application Research of Computers

基于小波倒谱系数和概率神经网络的取证说话人识别模型

Forensic speaker recognition model using wavelet cepstral coefficients and probabilistic neural network

免费全文下载 (已被下载 次)  
获取PDF全文
作者 雷磊,佘堃
机构 电子科技大学 信息与软件工程学院,成都 611731
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2018)04-0978-04
DOI 10.3969/j.issn.1001-3695.2018.04.004
摘要 取证说话人识别是一种利用犯罪现场留下的质疑语音来识别犯罪分子身份的技术。为了提高识别模型的抗噪能力,提出了基于小波倒谱系数(WCC)和概率神经网络(PNN)的取证说话人识别模型。该模型包含WCC特征提取和PNN分类两个步骤,WCC对噪声不敏感,所以使得该模型有抗噪能力;PNN是一种高效的分类算法,从而提高了模型识别性能。实验表明,该模型以提高时间消耗为代价提高了识别率和抗噪能力。
关键词 小波变换;概率神经网络;取证说话人识别
基金项目 四川省科技计划资助项目(2016GZ0073)
本文URL http://www.arocmag.com/article/01-2018-04-004.html
英文标题 Forensic speaker recognition model using wavelet cepstral coefficients and probabilistic neural network
作者英文名 Lei Lei, She Kun
机构英文名 SchoolofInformation&SoftwareEngineering,UniversityofElectronicScience&TechnologyofChina,Chengdu611731,China
英文摘要 Forensic speaker recognition refers to recognize the identity of the criminals from the questioned speech obtained by the criminal scene. To improve the anti-noise ability of recognition model, this paper proposed a forensic speaker recognition model based on wavelet cepstral coefficients(WCC) and probabilistic neural network(PNN). WCC was used to extract the feature vector that characterized the speech spectrum, because it was insensitive to noise. Moreover, PNN was a effective classification algorithm and was used as classifier to give out the forensic evidence in the proposed model. The experiments show that the proposed model obtains good performance in clear and noise environment at cost of increasing the time cost.
英文关键词 wavelet transform; probabilistic neural network; forensic speaker recognition
参考文献 查看稿件参考文献
  [1] Kinunen T, Li Haizhou. An overview of text-independent speaker reco-gnition:from features to supervectors[J] . Speech Communication, 2010, 52(1):12-40.
[2] Imai S. Cepstral analysis synthesis on the Mel frequency scale[C] //Proc of IEEE International Conference on Acoustics, Speech and Signal Processing. 1983:93-96.
[3] Ahmad K S, Thosar K J, Nirmal A S, et al. A unique approach in text-independent speaker recognition using MFCC feature sets and probabilistic neural network[C] //Proc of the 8th International Conference on Advances in Pattern Recognition. Piscataway:IEEE Press, 2015:1-6.
[4] Zhang Xueying, Bai Jing, Liang Wuzhou. The speech recognition system based on bark wavelet MFCC[C] //Proc of the 8th International Conference on Signal Processing. Piscataway:IEEE Press, 2006:16-20.
[5] Adam T B, Salam M S, Gunawan T S. Wavelet based cepstral coefficients for neural network speech recognition[C] //Proc of IEEE International Conference on Signal and Image Processing Application. Piscataway:IEEE Press, 2013:447-451.
[6] Reynolds D, Rose R. Robust text-independent speaker identification using Gaussian mixture speaker models[J] . Speech and Audio Processing, 1995, 3(1):72-83.
[7] 曹洁, 佘丽珍. 改进的说话人聚类初始化和GMM的多说话人识别[J] . 计算机应用研究, 2012, 29(2):590-593.
[8] Dash K, Padhi D, Panda B. Speaker identification using Mel frequency cepstral coefficient and BPNN[J] . International Journal of Advanced Research in Computer Science and Software Engineering, 2012, 2(4):327-332.
[9] 常飞, 乔欣, 张申, 等. 基于MFCC特征提取的故障预测与评估[J] . 计算机应用研究, 2015, 32(6):1716-1719.
[10] Arsikere H, Lulich S M. Estimating speaker height and subglottal resonance using MFCC and GMM[J] . Signal Processing Letters, 2014, 21(2):159-162.
[11] 侯建华, 陈稳, 刘欣达, 等. 基于异质性预矫正的小波域SAR图像去斑[J] . 计算机应用研究, 2016, 33(7):2224-2227.
[12] Ganchev T, Fakotakis N, Kokkinakis G. Text-independent speaker verification based on probabilistic neural network[C] //Proc of Acoustics. Piscataway:IEEE Press, 2002:159-166.
[13] Drygajlo A. From speaker recognition to forensic speaker recognition[C] //Proc of International Workshop on Biomedical Authentication. 2014:93-104.
[14] Biswas A, Sahu P K, Bhowmick A, et al. Feature extraction technique using ERB wavelet sub-band periodic and aperiodic decomposition for TIMIT phoneme recognition[J] . International Journal of Speech Technology, 2014, 17(4):389-399.
[15] Silveria M A, Schroeder C P, Costa D. Convolutive ICA-based forensic speaker identification using Mel frequency cepstral coefficients and Gaussian mixture model[J] . International Journal of Forensic Computer, 2013, 1:27-34.
收稿日期 2017/1/6
修回日期 2017/3/6
页码 978-981
中图分类号 TP391.4
文献标志码 A