Electronic apparatus and voice trigger method therefor

• US 10,147,444 B2
• Filed: 07/21/2016
• Issued: 12/04/2018
• Est. Priority Date: 11/03/2015
• Status: Active Grant

• Current Assignee: Airoha Technology Corporation (MediaTek, Inc.)
• Original Assignee: Airoha Technology Corporation (MediaTek, Inc.)
• Inventors: Wang, Chien-Chieh, Lin, Heng-Chih
• Primary Examiner(s): He, Jialong
• Application Number:
  US15/215,670
• Publication Number:
  US 20170125036A1
• Time in Patent Office: 866 Days
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 17/02 Preprocessing operations, e.g. segment selection; Pattern representation or modelling, e.g. based on linear discriminant analysis [LDA] or principal components; Feature selection or extraction
• CPC Class Codes: G10L 17/22 Interactive procedures; Man-machine interfaces
• CPC Class Codes: G10L 2015/0638 Interactive procedures
• CPC Class Codes: G10L 2015/088 Word spotting
• CPC Class Codes: G10L 2025/786 Adaptive threshold
• CPC Class Codes: G10L 25/78 Detection of presence or absence of voice signals switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M9/10

Abstract

A voice trigger method for an electronic apparatus is provided. The voice trigger method includes the following steps. Receive a current voice signal. Extract a current voice feature from the current voice signal. Determine whether a previous voice feature of a previous voice signal is stored. When the previous voice feature is stored, at least one of adjusting a confidence threshold and determining whether to wake up the electronic apparatus is performed according to whether the previous voice feature is corresponding to a false acceptance situation or a false rejection situation and the similarity between the current voice feature and the previous voice feature.

Claims

1. A method for waking up an electronic apparatus using voice trigger, comprising:

receiving a current voice signal;

extracting a current voice feature of the current voice signal;

determining whether a previous voice feature of a previous voice signal is stored;

wherein if it is determined that the previous voice feature is stored and a degree of similarity between the current voice feature and the previous voice feature is larger than or equivalent to a similarity threshold,

(A) adjusting a confidence threshold according to a false acceptance situation or a false rejection situation corresponding to the previous voice feature, or

(B) determining whether to wake up the electronic apparatus according to the false acceptance situation or the false rejection situation corresponding to the previous voice feature;

wherein if it is determined that the previous voice feature of the previous voice signal is not stored, then the method further comprises;

comparing the current voice feature with a key phrase voice model using a matching algorithm to obtain a confidence score;

determining whether the confidence score is larger than or equivalent to the confidence threshold;

waking up the electronic apparatus if the confidence score is larger than or equivalent to the confidence threshold; and

not waking up the electronic apparatus if the confidence score is smaller than the confidence threshold;

wherein after waking up the electronic apparatus, the method further comprises;

(1) determining whether the current voice feature corresponds to the false acceptance situation according to a user interaction and range of confidence score;

(2) if it is determined that the current voice feature corresponds to the false acceptance situation, then the current voice feature is stored and the confidence threshold is adjusted according to the user interaction;

wherein the user interaction comprises a frequency of consecutive false acceptance, and the step of adjusting the confidence threshold according to the user interaction comprises;

(1) determining whether the frequency of consecutive false acceptance is larger than a frequency threshold; and

(2) increasing the confidence threshold if the frequency of consecutive false acceptance is larger than the frequency threshold.

2. The method according to claim 1, wherein if it is determined that the previous voice feature corresponds to the false acceptance situation and the degree of similarity is larger than or equivalent to the similarity threshold, then the confidence threshold is increased or the electronic apparatus is not woken up.

3. The method according to claim 1, wherein if it is determined that the previous voice feature corresponds to the false rejection situation and the degree of similarity is larger than or equivalent to the similarity threshold, then the confidence threshold is lowered or the electronic apparatus is woken up.

4. The method according to claim 1, wherein after not waking up the electronic apparatus, the method further comprises:

determining whether the current voice feature corresponds to a false rejection situation according to the user interaction and range of confidence score;

wherein if it is determined that the current voice feature corresponds to the false rejection situation, then the current voice feature is stored and the confidence threshold is adjusted according to the user interaction.

5. The method according to claim 4, herein the user interaction comprises a frequency of consecutive false rejection, and the step of adjusting the confidence threshold according to the user interaction comprises:

determining whether the frequency of consecutive false rejection is larger than the frequency threshold;

lowering the confidence threshold if the frequency of consecutive false rejection is larger than the frequency threshold.

6. An electronic apparatus, comprising:

a storage device;

an voice receiver used for receiving a current voice signal;

a processor used for extracting a current voice feature of the current voice signal and determining whether a previous voice feature of a previous voice signal is stored in the storage device;

wherein if it is determined that the previous voice feature is stored in the storage device and a degree of similarity between the current voice feature and the previous voice feature is larger than or equivalent to a similarity threshold, then

(A) the processor adjusts a confidence threshold according to a false acceptance situation or a false rejection situation corresponding to the previous voice feature, or

(B) the processor determines whether to wake up the electronic apparatus according to the false acceptance situation or the false rejection situation corresponding to the previous voice feature;

wherein if it is determined that the previous voice feature of the previous voice signal is not stored, then

(1) the processor compares the current voice feature with a key phrase voice model using a matching algorithm to obtain a confidence score and further determines whether the confidence score is larger than or equivalent to the confidence threshold;

(2) the processor wakes up the electronic apparatus if the confidence score is larger than or equivalent to the confidence threshold, but does not wake up the electronic apparatus if the confidence score is smaller than the confidence threshold;

wherein after waking up the electronic apparatus, the processor further determines whether the current voice feature corresponds to the false acceptance situation according to a user interaction and range of confidence score;

if it is determined that the current voice feature corresponds to the false acceptance situation, then the processor stores the current voice feature to the storage device and adjusts the confidence threshold according to the user interaction;

wherein the user interaction comprises a frequency of consecutive false acceptance, the processor further determines whether the frequency of consecutive false acceptance is larger than a frequency threshold;

the confidence threshold is increased if it is determined that the frequency of consecutive false acceptance is larger than the frequency threshold.

7. The electronic apparatus according to claim 6, wherein the processor increases the confidence threshold or does not wake up the electronic apparatus if it is determined that the previous voice feature corresponds to the false acceptance situation and the degree of similarity is larger than or equivalent to the similarity threshold.

8. The electronic apparatus according to claim 6, wherein the processor decrease the confidence threshold or wakes up the electronic apparatus if it is determined that the previous voice feature corresponds to the false rejection situation and the degree of similarity is larger than or equivalent to the similarity threshold.

9. The electronic apparatus according to claim 6, wherein after not waking up the electronic apparatus, the processor further determines whether the current voice feature corresponds to a false rejection situation according to the user interaction and range of confidence score;

if it is determined that the current voice feature corresponds to the false rejection situation, then the processor stores the current voice feature to the storage device and adjusts the confidence threshold according to the user interaction.

10. The electronic apparatus according to claim 9, wherein the user interaction comprises a frequency of consecutive false rejection, the processor further determines whether the frequency of consecutive false rejection is larger than the frequency threshold;

the confidence threshold is lowered if it is determined that the frequency of consecutive false rejection is larger than the frequency threshold.