Spoofing speech detection using temporal convolutional neural network
Spoofing speech detection aims to differentiate spoofing speech from natural speech. Frame-based features are usually used in most of previous works. Although multiple frames or dynamic features are used to form a super-vector to represent the temporal information, the time span covered by these fea...
| Egile Nagusiak: | , , , |
|---|---|
| Beste egile batzuk: | |
| Formatua: | Conference Paper |
| Hizkuntza: | English |
| Argitaratua: |
2018
|
| Gaiak: | |
| Sarrera elektronikoa: | https://hdl.handle.net/10356/89639 http://hdl.handle.net/10220/47064 |
| _version_ | 1826130578410307584 |
|---|---|
| author | Xiao, Xiong Li, Haizhou Tian, Xiaohai Chng, Eng Siong |
| author2 | School of Computer Science and Engineering |
| author_facet | School of Computer Science and Engineering Xiao, Xiong Li, Haizhou Tian, Xiaohai Chng, Eng Siong |
| author_sort | Xiao, Xiong |
| collection | NTU |
| description | Spoofing speech detection aims to differentiate spoofing speech from natural speech. Frame-based features are usually used in most of previous works. Although multiple frames or dynamic features are used to form a super-vector to represent the temporal information, the time span covered by these features are not sufficient. Most of the systems failed to detect the non-vocoder or unit selection based spoofing attacks. In this work, we propose to use a temporal convolutional neural network (CNN) based classifier for spoofing speech detection. The temporal CNN first convolves the feature trajectories with a set of filters, then extract the maximum responses of these filters within a time window using a max-pooling layer. Due to the use of max-pooling, we can extract useful information from a long temporal span without concatenating a large number of neighbouring frames, as in feedforward deep neural network (DNN). Five types of feature are employed to access the performance of proposed classifier. Experimental results on ASVspoof 2015 corpus show that the temporal CNN based classifier is effective for synthetic speech detection. Specifically, the proposed method brings a significant performance boost for the unit selection based spoofing speech detection. |
| first_indexed | 2024-10-01T07:58:39Z |
| format | Conference Paper |
| id | ntu-10356/89639 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T07:58:39Z |
| publishDate | 2018 |
| record_format | dspace |
| spelling | ntu-10356/896392020-03-07T11:48:46Z Spoofing speech detection using temporal convolutional neural network Xiao, Xiong Li, Haizhou Tian, Xiaohai Chng, Eng Siong School of Computer Science and Engineering 2016 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA) NTU-UBC Research Centre of Excellence in Active Living for the Elderly Temasek Laboratories DRNTU::Engineering::Computer science and engineering Convolutional Neural Network (CNN) Speech Detection Spoofing speech detection aims to differentiate spoofing speech from natural speech. Frame-based features are usually used in most of previous works. Although multiple frames or dynamic features are used to form a super-vector to represent the temporal information, the time span covered by these features are not sufficient. Most of the systems failed to detect the non-vocoder or unit selection based spoofing attacks. In this work, we propose to use a temporal convolutional neural network (CNN) based classifier for spoofing speech detection. The temporal CNN first convolves the feature trajectories with a set of filters, then extract the maximum responses of these filters within a time window using a max-pooling layer. Due to the use of max-pooling, we can extract useful information from a long temporal span without concatenating a large number of neighbouring frames, as in feedforward deep neural network (DNN). Five types of feature are employed to access the performance of proposed classifier. Experimental results on ASVspoof 2015 corpus show that the temporal CNN based classifier is effective for synthetic speech detection. Specifically, the proposed method brings a significant performance boost for the unit selection based spoofing speech detection. NRF (Natl Research Foundation, S’pore) Accepted version 2018-12-18T07:45:21Z 2019-12-06T17:30:03Z 2018-12-18T07:45:21Z 2019-12-06T17:30:03Z 2016-12-01 2016 Conference Paper Tian, X., Xiao, X., Chng, E. S., & Li, H. (2016). Spoofing speech detection using temporal convolutional neural network. 2016 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA). doi:10.1109/APSIPA.2016.7820738 https://hdl.handle.net/10356/89639 http://hdl.handle.net/10220/47064 10.1109/APSIPA.2016.7820738 200465 en © 2016 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The published version is available at: [http://dx.doi.org/10.1109/APSIPA.2016.7820738]. 6 p. application/pdf |
| spellingShingle | DRNTU::Engineering::Computer science and engineering Convolutional Neural Network (CNN) Speech Detection Xiao, Xiong Li, Haizhou Tian, Xiaohai Chng, Eng Siong Spoofing speech detection using temporal convolutional neural network |
| title | Spoofing speech detection using temporal convolutional neural network |
| title_full | Spoofing speech detection using temporal convolutional neural network |
| title_fullStr | Spoofing speech detection using temporal convolutional neural network |
| title_full_unstemmed | Spoofing speech detection using temporal convolutional neural network |
| title_short | Spoofing speech detection using temporal convolutional neural network |
| title_sort | spoofing speech detection using temporal convolutional neural network |
| topic | DRNTU::Engineering::Computer science and engineering Convolutional Neural Network (CNN) Speech Detection |
| url | https://hdl.handle.net/10356/89639 http://hdl.handle.net/10220/47064 |
| work_keys_str_mv | AT xiaoxiong spoofingspeechdetectionusingtemporalconvolutionalneuralnetwork AT lihaizhou spoofingspeechdetectionusingtemporalconvolutionalneuralnetwork AT tianxiaohai spoofingspeechdetectionusingtemporalconvolutionalneuralnetwork AT chngengsiong spoofingspeechdetectionusingtemporalconvolutionalneuralnetwork |