US 10,186,254 B2Granted

Context-based endpoint detection

AssigneeApple Inc.

InventorsShaun E. Williams, Henry G. Mason, Mahesh Krishnamoorthy, Matthias Paulik, Neha Agrawal, Sachin S. Kajarekar, Selen Uguroglu, Ali S. Mohamed

Life of this patent

Estimating the term…

patentgrep

Loading patent…

Abstract

The present disclosure generally relates to context-based endpoint detection in user speech input. A method for identifying an endpoint of a spoken request by a user may include receiving user input of natural language speech including one or more words; identifying at least one context associated with the user input; generating a probability, based on the at least one context associated with the user input, that a location in the user input is an endpoint; determining whether the probability is greater than a threshold; and in accordance with a determination that the probability is greater than the threshold, identifying the location in the user input as the endpoint.

Claims66

Dependent claims

A non-transitory computer-readable storage medium storing one or more programs, the one or more programs comprising instructions, which when executed by an electronic device, cause the electronic device to:

receive user input comprising natural language speech including one or more words;
identify at least one context associated with the user input;
divide the user input into a plurality of audio frames;
determine whether each frame of the plurality of frames includes audio information associated with the user input;
in accordance with a determination that a frame includes audio information associated with the user input, determine whether a threshold number of frames of silence follow;
in accordance with a determination that the threshold number of frames of silence follow, cease recording the user input;
in accordance with a determination that the threshold number of frames of silence do not follow, continue recording the user input;
generate a probability, based on the at least one context associated with the user input, that a location in the user input is an endpoint;
determine whether the probability is greater than a threshold;
and
in accordance with a determination that the probability is greater than the threshold, identify the location in the user input as the endpoint.

8 more elements

A method for identifying an endpoint of a spoken request by a user, comprising:

at a device with one or more processors and memory:
receiving user input comprising natural language speech including one or more words;
identifying at least one context associated with the user input;
dividing the user input into a plurality of audio frames;
determining whether each frame of the plurality of frames includes audio information associated with the user input;
in accordance with a determination that a frame includes audio information associated with the user input, determining whether a threshold number of frames of silence follow;
in accordance with a determination that the threshold number of frames of silence follow, ceasing to record the user input;
in accordance with a determination that the threshold number of frames of silence do not follow, continuing to record the user input;
generating a probability, based on the at least one context associated with the user input, that a location in the user input is an endpoint;
determining whether the probability is greater than a threshold;
and
in accordance with a determination that the probability is greater than the threshold, identifying the location in the user input as the endpoint.

9 more elements

An electronic device, comprising:

a display;
a memory;
a processor coupled to the display and the memory;
and programs stored in the memory to be executed by the processor, the programs comprising instructions for:
receiving user input comprising natural language speech including one or more words;
identifying at least one context associated with the user input;
dividing the user input into a plurality of audio frames;
determining whether each frame of the plurality of frames includes audio information associated with the user input;
in accordance with a determination that a frame includes audio information associated with the user input, determining whether a threshold number of frames of silence follow;
in accordance with a determination that the threshold number of frames of silence follow, ceasing to record the user input;
in accordance with a determination that the threshold number of frames of silence do not follow, continuing to record the user input;
generating a probability, based on the at least one context associated with the user input, that a location in the user input is an endpoint;
determining whether the probability is greater than a threshold;
and
in accordance with a determination that the probability is greater than the threshold, identifying the location in the user input as the endpoint.

12 more elements

Dependencies are inferred from the claim text.

Facts

Application: 14/846,667Filed 4 Sep 2015 · us provisional application
Priority: 7 Jun 2015
Published: US 2016/0358598 A18 Dec 2016
Granted: 22 Jan 2019
Term ends: Not estimatedThe source records don’t support an estimate.

Classification5 codes

G10L 15/04
Speech recognition: segmentation; Word boundary detection
G10L 17/02
Speaker identification or verification techniques: preprocessing operations, e.g. segment selection; Pattern representation or modelling, e.g.…
G10L 25/87
Detection of presence or absence of voice signals: detection of discrete points within a voice signal
G10L 2025/783
Detection of presence or absence of voice signals: based on threshold decision