Name
Abstract | ||
|---|---|---|
Automatic speaker verification (ASV) systems are highly vulnerable against spoofing attacks, also known as imposture. With recent developments in speech synthesis and voice conversion technology, it has become important to detect synthesized or voice-converted speech for the security of ASV systems. In this paper, we compare five different classifiers used in speaker recognition to detect synthetic speech. Experimental results conducted on the ASVspoof 2015 dataset show that support vector machines with generalized linear discriminant kernel (GLDS-SVM) yield the best performance on the development set with the EER of 0.12 % whereas Gaussian mixture model (GMM) trained using maximum likelihood (ML) criterion with the EER of 3.01 % is superior for the evaluation set. |
Year | Venue | Keywords |
|---|---|---|
2015 | 16TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION (INTERSPEECH 2015), VOLS 1-5 | spoof detection, countermeasures, speaker recognition |
Field | DocType | Citations |
Kernel (linear algebra),Speech synthesis,Pattern recognition,Spoofing attack,Computer science,Voice activity detection,Support vector machine,Speech recognition,Speaker recognition,Artificial intelligence,Linear discriminant analysis,Mixture model | Conference | 13 |
PageRank | References | Authors |
0.56 | 15 | 4 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Cemal Hanilçi | 1 | 171 | 11.23 |
| Tomi Kinnunen | 2 | 1323 | 86.67 |
| Md. Sahidullah | 3 | 326 | 24.99 |
| Aleksandr Sizov | 4 | 96 | 4.54 |