Sixth European Conference on Speech Communication and Technology
(EUROSPEECH'99)

Budapest, Hungary
September 5-9, 1999

A Language-Independent Personal Voice Controller with Embedded Speaker Verification

Qi Li, Augustine Tsai, Weon-Goo Kim

Multimedia Communications Research Laboratory, Bell Labs, Lucent Technologies Murray Hill, NJ, USA

In this paper, we introduce a personal voice controller for name dialing. As a personal system, all the control commands are trained based on owner-selected phrases in the owner's language. Since the owner's voice characteristics is modeled together with the command voice, the system has embedded speaker verification capability without additional processing. In other words, once the system is trained, the system can only accept the voice commands from the owner. Impostor's command can be rejected although the command may be in the owner's list. The controller can recognize the speaker's command and dial the destinate phone number automatically. Such a system is useful for wireless handsets and portable communication devices. Its implementation requires much less memory space and computation resource compared to a speaker-independent system. Preliminary experiments showed that the proposed system has the state-of-the-art performances in both speech recognition and speaker verification.


Full Paper (PDF)   Gnu-Zipped Postscript

Bibliographic reference.  Li, Qi / Tsai, Augustine / Kim, Weon-Goo (1999): "A language-independent personal voice controller with embedded speaker verification", In EUROSPEECH'99, 1207-1210.