Performance enhancement of speaker identification systems using speech encryption and cancelable features
| Autor: | Naglaa F. Soliman, Mahmoud I. Abdalla, Fathi E. Abd El-Samie, Zhraa Mostfa |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Linguistics and Language
Biometrics Artificial neural network Computer science business.industry Speech recognition 020208 electrical & electronic engineering Feature extraction 020206 networking & telecommunications 02 engineering and technology Speaker recognition Encryption Language and Linguistics Human-Computer Interaction 0202 electrical engineering electronic engineering information engineering Cryptosystem Wireless Computer Vision and Pattern Recognition Noise (video) business Software |
| Zdroj: | International Journal of Speech Technology. 20:977-1004 |
| ISSN: | 1572-8110 1381-2416 |
| DOI: | 10.1007/s10772-017-9435-z |
| Popis: | Biometric systems based on speech features constitute a new and evolving trend in security. This paper is concerned with speaker identification systems. It studies traditional speaker identification systems based on cepstral analysis and neural classification. The paper develops the idea of remote access systems with speaker identification concepts by introducing efficient cryptosystems to achieve a large degree of security in these remote access speaker identification systems. The proposed approaches depend on chaos theory to maintain a low sensitivity to noise effect. Moreover, the concepts of cancelable biometrics are developed in this paper for more secure speaker identification. As known in the literature, cancelable image biometrics are used to save the features of the users from being stolen. If a similar approach is adopted in wireless access speaker identification systems as in this paper, the security can be enhanced to a great extent. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink