Efficient cancelable speaker identification system based on a hybrid structure of DWT and SVD
- Resource Type
- Original Paper
- Authors
- Abdelwahab, Khaled M.; El-atty, Saied Abd; Brisha, Ayman M.; Abd El-Samie, Fathi E.
- Source
- International Journal of Speech Technology. 25(1):279-288
- Subject
- Audio processing
Biometrics
Cancelable biometrics
Pattern recognition
Security
Speaker identification
- Language
- English
- ISSN
- 1381-2416
1572-8110
Traditional biometric systems are subjected to several attacks. If the original database of human biometrics is compromised, the biometric data is lost forever. This paper presents a sophisticated implementation of one of the watermarking algorithms in the field of cancelable speaker identification. A watermark strength factor is used to control the level of intended distortion created in the speech signal prior to the identification process. The practical simulation of the proposed system proves that it is possible through a watermarking algorithm to distort speech signals intentionally to generate cancelable speech templates.