摘要 |
In a voice-monitoring system (eg. a mobile device with a microphone), a Speech Recognition (SR) algorithm identifies words in the input signal 32 and uses some of these words 33 (eg. a name after an initial hello) to identify a person involved in the conversation 35. Portions of their voice are then used to generate an audio signature 36 (via eg. their pronunciation, timing or frequency) which may be linked to the detected identity, stored 37 and displayed to a user. |