Artificial intelligence apparatus
AssigneeLG Electronics Inc.
InventorHansuk Shim
Life of this patent
Estimating the term…
Abstract
Disclosed herein is an artificial intelligence apparatus including an input interface configured to receive speech data, and a processor configured to detect a non-utterance interval included in the speech data and determine presence/absence of a second utterance after the non-utterance interval according to characteristics of a first utterance before the non-utterance interval, when the non-utterance interval exceeds a set time.
Claims19
An artificial intelligence apparatus comprising:
- at least one sensor configured to obtain audio data;
- and
- one or more processors configured to:
- obtain a wakeup word, via the at least one sensor, to activate speech recognition;
- based on activating the speech recognition, detect a non-utterance interval in the audio data;
- and
- based on a length of the non-utterance interval exceeding a threshold length:
- extract a first utterance preceding the non-utterance interval in the audio data;
- determine whether a second utterance follows the non-utterance interval based on one or more characteristics of the first utterance by inputting the first utterance to an artificial intelligence model for extracting one or more speech characteristics and one or more context characteristics of the first utterance;
- and
- maintain the activated speech recognition, upon determining that the second utterance follows the non-utterance interval.
A method of operating an artificial intelligence apparatus, the method comprising:
- obtaining audio data;
- obtaining a wakeup word to activate speech recognition;
- based on activating the speech recognition, detecting a non-utterance interval in the audio data;
- and
- based on a length of the non-utterance interval exceeding a threshold length:
- extracting a first utterance preceding the non-utterance interval in the audio data;
- determining whether a second utterance follows the non-utterance interval based on one or more characteristics of the first utterance by inputting the first utterance to an artificial intelligence model for extracting one or more speech characteristics and one or more context characteristics of the first utterance;
- and
- maintaining the activated speech recognition, upon determining that the second utterance follows the non-utterance interval.
A non-transitory computer-readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to:
- obtain audio data;
- obtain a wakeup word to activate speech recognition;
- based on activating the speech recognition, detect a non-utterance interval in the audio data;
- and
- based on a length of the non-utterance interval exceeding a threshold length:
- extract a first utterance preceding the non-utterance interval in the audio data;
- determine whether a second utterance follows the non-utterance interval based on one or more characteristics of the first utterance by inputting the first utterance to an artificial intelligence model for extracting one or more speech characteristics and one or more context characteristics of the first utterance;
- and
- maintain the activated speech recognition, upon determining that the second utterance follows the non-utterance interval.
Dependencies are inferred from the claim text.