Track: Keynote VoiceTech |
| Trust in Voice. Verification in the Age of Synthetic Speech: Is it live -- or is it….? |
| As machine learning improves our ability to clone and synthesize voices that sound increasingly human, how do we ensure that we know who (or what) we are talking to? In this presentation we’ll look at the challenges of establishing “liveness” and the extent to which we can use voice biometrics to identify a speaker, primarily over the phone system, where channel characteristics and call quality impose additional barriers to correctly scoring a call as a live human and correctly identifying the caller. Perhaps surprisingly, voice analysis can be more difficult than video analysis, for which we have more information and the ability to interact with the image. |
|
| Presentation Video |
| Presentation Notes |
|
Parkinson-Trudt-in-Voice-2.pdf |
Real Time Communications Conference & Expo at Illinois Tech
IEEE International Conference
Follow Us
Share This
