EI / SCOPUS / CSCD 收录

中文核心期刊

WANG Liwei, WANG Zuoying. Structural risk minimization principle based gaussian mixture modeling[J]. ACTA ACUSTICA, 2003, 28(5): 403-408. DOI: 10.15949/j.cnki.0371-0025.2003.05.004
Citation: WANG Liwei, WANG Zuoying. Structural risk minimization principle based gaussian mixture modeling[J]. ACTA ACUSTICA, 2003, 28(5): 403-408. DOI: 10.15949/j.cnki.0371-0025.2003.05.004

Structural risk minimization principle based gaussian mixture modeling

More Information
  • PACS: 
    • 43.70  (Speech production)
  • Received Date: December 05, 2001
  • Revised Date: February 27, 2002
  • Available Online: August 03, 2022
  • Gaussian mixture model are commonly used as probability density of observed data in Hidden Markov Models. The parameters of the Gaussian mixture model is estimated based on Maximum Likelihood principle in almost all automatic speech recognition systems. However, the value of the likelihood would monotonically increase with the increases of the number of the components. On the other hand, parameters of over-complex model could not be welltrained, and give bad results for testing data. In this paper, we present an estimation of the parameters of Gaussian mixture models based on Structural Risk Minimization principle which is the key principle of statistical learning theory. We further generalized the method for the purpose that simple model is more attractive in real-time systems. The new methodology could reduce the error rate consistently compared with baseline systems which has the same average number of components.
  • Related Articles

    [1]WANG Ran, WANG Guangzhe, ZHANG Chenyu, GUO Qixin, ZHANG Yongli, YU Liang, GAO Yuan, CHEN Nuo. Gaussian mixture model fitting and suppression of towed array flow noise[J]. ACTA ACUSTICA, 2024, 49(5): 1030-1040. DOI: 10.12395/0371-0025.2023033
    [2]LU Da, DUAN Rui, YANG Kunde. Tensor decomposition based normal mode parameter estimation[J]. ACTA ACUSTICA, 2023, 48(4): 743-760. DOI: 10.15949/j.cnki.0371-0025.2023.04.028
    [3]SHI Qian, CHEN Hangting, ZHANG Pengyuan. Direction-guided speech enhancement method of the spatial mixture model[J]. ACTA ACUSTICA, 2022, 47(1): 139-150. DOI: 10.15949/j.cnki.0371-0025.2022.01.016
    [4]XIA Ping, ZHANG Guangyi, LEI Bangjun, GONG Guoqiang, ZOU Yaobin, TANG Tinglong. Sonar image enhancement of digraph and Gaussian mixture model in complex contourlet domain[J]. ACTA ACUSTICA, 2021, 46(4): 529-539. DOI: 10.15949/j.cnki.0371-0025.2021.04.005
    [5]SONG Guoli, GUO Xinyi, MA Li. Study on the non-Gaussian ocean ambient noise model with certain kurtosis[J]. ACTA ACUSTICA, 2019, 44(5): 887-896. DOI: 10.15949/j.cnki.0371-0025.2019.05.009
    [6]HOU Leijing, GUO Tingting, SUN Yan, QI Yingjie, YING Dongwen, TANG Min, YAN Yonghong. A personalized Gaussian mixture model modeling method for heart sound segmentation[J]. ACTA ACUSTICA, 2019, 44(1): 20-27. DOI: 10.15949/j.cnki.0371-0025.2019.01.003
    [7]ZHANG Chi, MA Xiaochuan, WU Yongqing, ZHANG Shuhao, ZHAN Fei. A method of estimating the clutter suppressing performance of active sonar waveform based on the reverberation mixture distribution model[J]. ACTA ACUSTICA, 2018, 43(4): 565-574. DOI: 10.15949/j.cnki.0371-0025.2018.04.016
    [8]XU Chundong, ZHANG Zhen, ZHAN Ge, YING Dongwen, LI Junfeng, YAN Yonghong. Noise power estimation based on constrained sequential Gaussian mixture model for speech enhancement[J]. ACTA ACUSTICA, 2017, 42(5): 633-640. DOI: 10.15949/j.cnki.0371-0025.2017.05.015
    [9]LI Yangchun, YU Yibiao. Voice conversion using structured Gaussian mixture model in eigen space[J]. ACTA ACUSTICA, 2015, 40(1): 12-19. DOI: 10.15949/j.cnki.0371-0025.2015.01.002
    [10]ZHANG Yong, HU Ruimin. Speech wideband extension based on Gaussian mixture model[J]. ACTA ACUSTICA, 2009, 34(5): 471-480. DOI: 10.15949/j.cnki.0371-0025.2009.05.004

Catalog

    Article Metrics

    Article views (40) PDF downloads (3) Cited by()
    Related

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return