DNN Filter Bank Cepstral Coefficients for Spoofing Detection
With the development of speech synthesis techniques, automatic speaker verification systems face the serious challenge of spoofing attack. In order to improve the reliability of speaker verification systems, we develop a new filter bank-based cepstral feature, deep neural network (DNN) filter bank c...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2017-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/7886361/ |
| _version_ | 1818644534270099456 |
|---|---|
| author | Hong Yu Zheng-Hua Tan Yiming Zhang Zhanyu Ma Jun Guo |
| author_facet | Hong Yu Zheng-Hua Tan Yiming Zhang Zhanyu Ma Jun Guo |
| author_sort | Hong Yu |
| collection | DOAJ |
| description | With the development of speech synthesis techniques, automatic speaker verification systems face the serious challenge of spoofing attack. In order to improve the reliability of speaker verification systems, we develop a new filter bank-based cepstral feature, deep neural network (DNN) filter bank cepstral coefficients, to distinguish between natural and spoofed speech. The DNN filter bank is automatically generated by training a filter bank neural network (FBNN) using natural and synthetic speech. By adding restrictions on the training rules, the learned weight matrix of FBNN is band limited and sorted by frequency, similar to the normal filter bank. Unlike the manually designed filter bank, the learned filter bank has different filter shapes in different channels, which can capture the differences between natural and synthetic speech more effectively. The experimental results on the ASVspoof 2015 database show that the Gaussian mixture model maximum-likelihood classifier trained by the new feature performs better than the state-of-the-art linear frequency triangle filter bank cepstral coefficients-based classifier, especially on detecting unknown attacks. |
| first_indexed | 2024-12-17T00:16:22Z |
| format | Article |
| id | doaj.art-1a137fcec578423da0236f57ca436f73 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-17T00:16:22Z |
| publishDate | 2017-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-1a137fcec578423da0236f57ca436f732022-12-21T22:10:42ZengIEEEIEEE Access2169-35362017-01-0154779478710.1109/ACCESS.2017.26870417886361DNN Filter Bank Cepstral Coefficients for Spoofing DetectionHong Yu0https://orcid.org/0000-0002-7868-5201Zheng-Hua Tan1Yiming Zhang2Zhanyu Ma3https://orcid.org/0000-0003-2950-2488Jun Guo4Pattern Recognition and Intelligent System Laboratory, Beijing University of Posts and Telecommunications, Beijing, ChinaDepartment of Electronic Systems, Aalborg University, Aalborg, DenmarkInternational School, Beijing University of Posts and Telecommunications, Beijing, ChinaPattern Recognition and Intelligent System Laboratory, Beijing University of Posts and Telecommunications, Beijing, ChinaPattern Recognition and Intelligent System Laboratory, Beijing University of Posts and Telecommunications, Beijing, ChinaWith the development of speech synthesis techniques, automatic speaker verification systems face the serious challenge of spoofing attack. In order to improve the reliability of speaker verification systems, we develop a new filter bank-based cepstral feature, deep neural network (DNN) filter bank cepstral coefficients, to distinguish between natural and spoofed speech. The DNN filter bank is automatically generated by training a filter bank neural network (FBNN) using natural and synthetic speech. By adding restrictions on the training rules, the learned weight matrix of FBNN is band limited and sorted by frequency, similar to the normal filter bank. Unlike the manually designed filter bank, the learned filter bank has different filter shapes in different channels, which can capture the differences between natural and synthetic speech more effectively. The experimental results on the ASVspoof 2015 database show that the Gaussian mixture model maximum-likelihood classifier trained by the new feature performs better than the state-of-the-art linear frequency triangle filter bank cepstral coefficients-based classifier, especially on detecting unknown attacks.https://ieeexplore.ieee.org/document/7886361/Speaker verificationspoofing detectionDNN filter bank cepstral coefficientsfilter bankneural network |
| spellingShingle | Hong Yu Zheng-Hua Tan Yiming Zhang Zhanyu Ma Jun Guo DNN Filter Bank Cepstral Coefficients for Spoofing Detection IEEE Access Speaker verification spoofing detection DNN filter bank cepstral coefficients filter bank neural network |
| title | DNN Filter Bank Cepstral Coefficients for Spoofing Detection |
| title_full | DNN Filter Bank Cepstral Coefficients for Spoofing Detection |
| title_fullStr | DNN Filter Bank Cepstral Coefficients for Spoofing Detection |
| title_full_unstemmed | DNN Filter Bank Cepstral Coefficients for Spoofing Detection |
| title_short | DNN Filter Bank Cepstral Coefficients for Spoofing Detection |
| title_sort | dnn filter bank cepstral coefficients for spoofing detection |
| topic | Speaker verification spoofing detection DNN filter bank cepstral coefficients filter bank neural network |
| url | https://ieeexplore.ieee.org/document/7886361/ |
| work_keys_str_mv | AT hongyu dnnfilterbankcepstralcoefficientsforspoofingdetection AT zhenghuatan dnnfilterbankcepstralcoefficientsforspoofingdetection AT yimingzhang dnnfilterbankcepstralcoefficientsforspoofingdetection AT zhanyuma dnnfilterbankcepstralcoefficientsforspoofingdetection AT junguo dnnfilterbankcepstralcoefficientsforspoofingdetection |