Back to results
US 11,699,433 B2Granted
Dynamic wakeword detection
AssigneeAmazon Technologies, Inc.
InventorsGengshen Fu, Shiv Naga Prasad Vitaladevuni, Paul McIntyre, Shuang Wu
Life of this patent
Estimating the term…
Abstract
Techniques for using a dynamic wakeword detection threshold are described. A device detects a wakeword in audio data using a first wakeword detection threshold value. Thereafter, the device receives audio including speech. If the device receives the audio within a predetermined duration of time after detecting the previous wakeword, the device attempts to detect a wakeword in second audio data, corresponding to the audio including the speech, using a second, lower wakeword detection threshold value.
Claims22
A computer-implemented method comprising:
- receiving first audio data corresponding to a first utterance;
- using a first component operating at a first setting to determine the first audio data includes a first representation of a wakeword;
- based at least in part on the first audio data including the first representation of the wakeword, causing further processing to be performed using the first audio data;
- determining, based at least in part on a result of the further processing, a likelihood of detecting a second utterance within a duration of time of the first utterance, wherein the second utterance is a different utterance from the first utterance;
- prior to detection of the second utterance and based at least in part on the likelihood of detecting the second utterance, configuring the first component to operate using a second setting different from the first setting;
- receiving second audio data;
- and
- using the first component operating at the second setting to determine whether the second audio data includes a second representation of the wakeword.
5 more elements
A system comprising:
- at least one processor;
- and
- at least one memory comprising instructions that, when executed by the at least one processor, cause the system to:
- receive first audio data corresponding to a first utterance;
- use a first component operating at a first setting to determine the first audio data includes a first representation of a wakeword;
- based at least in part on the first audio data including the first representation of the wakeword, cause further processing to be performed using the first audio data;
- determine, based at least in part on a result of the further processing, a likelihood of detecting a second utterance within a duration of time of the first utterance, wherein the second utterance is a different utterance from the first utterance;
- prior to detection of the second utterance and based at least in part on the likelihood of detecting the second utterance, configure the first component to operate using a second setting different from the first setting;
- receive second audio data;
- and
- use the first component operating at the second setting to determine whether the second audio data includes a second representation of the wakeword.
8 more elements
Dependencies are inferred from the claim text.