Wake-up method, device and equipment for intelligent equipment

A smart device and sample technology, applied in the direction of program control device, program control design, voice analysis, etc., can solve the problems of high resource occupation and poor wake-up effect, and achieve the effect of reducing dependence, improving performance, and reducing occupied resources

Pending Publication Date: 2021-06-01
SOUNDAI TECH CO LTD
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The present disclosure provides a wake-up method, device, and device for smart devices, which solve the problems of the smart device wake-up schemes in related technologies that occupy too many resources and rely on alignment processing for poor wake-up effects.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Wake-up method, device and equipment for intelligent equipment
  • Wake-up method, device and equipment for intelligent equipment
  • Wake-up method, device and equipment for intelligent equipment

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0126] An embodiment of the present disclosure provides a flowchart of a method for waking up a smart device, such as figure 2 shown, including:

[0127] Step S201, obtaining an audio data sequence;

[0128] Because in the solution of the embodiment of the present disclosure, the label predicted by the following wake-up model is the label corresponding to each audio data sequence, if the down-sampling process is performed, the wake-up result will not be affected, therefore, the above-mentioned audio data sequence can be down-sampled Processing, in order to achieve the effect of greatly reducing the amount of calculation.

[0129] As an optional implementation, after obtaining the audio data sequence, it also includes:

[0130] The above audio data sequence is down-sampled.

[0131] Wherein, the ratio of downsampling is specifically set according to specific implementation conditions.

[0132] Step S202, input the audio data sequence into the wake-up model, use the wake-up...

Embodiment approach 1

[0138] Embodiment 1: The sequence label is a sequence label that recognizes the wake-up sentence sequence or does not recognize the wake-up sentence sequence.

[0139] The specific manner of determining the sequence label in the above-mentioned embodiment 1 is as described in the above-mentioned step S102, and will not be repeated here.

[0140] It should be noted that, the content of the above-mentioned sequence label can be a substitute symbol for identifying a wake-up sentence sequence or not recognizing a wake-up sentence sequence. It can also be the specific content of the recognized sequence, for example, lan se ping guo.

Embodiment approach 2

[0141] Embodiment 2: The sequence labels are sequence labels and probabilities of recognized wake-up sentence sequences and sequence labels and probabilities of unrecognized wake-up sentence sequences.

[0142] According to the sequence characteristics of each wake-up statement sample in the forward sample set and the above-mentioned first sequence characteristics, predict and identify the first sequence label and first probability of the wake-up statement sequence;

[0143] According to the sequence characteristics of each non-awakening sentence sample in the reverse sample set and the above-mentioned first sequence characteristics, the second sequence label and the second probability of the wake-up sentence sequence are not recognized in the prediction mark;

[0144] According to the first sequence label and the first probability, the second sequence label and the second probability, predict whether the wake-up sentence sequence is recognized.

[0145] It should be noted tha...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention provides a wake-up method, device and equipment for intelligent equipment. The method comprises the following steps: acquiring an audio data sequence; inputting the audio data sequence into a wake-up model, performing feature extraction on the audio data sequence by using the wake-up model to obtain a first sequence feature, and predicting whether a wake-up statement sequence is identified and outputting a corresponding sequence tag according to a second sequence feature and the first sequence feature of a wake-up statement sample used for waking up the intelligent device; and when it is determined that the sequence tag of the identified wake-up statement sequence is output, waking up the intelligent device. According to the invention, resources occupied by the wake-up function of the intelligent device can be reduced, the dependence of the wake-up function on alignment processing is reduced, and the wake-up effect is improved.

Description

technical field [0001] The present disclosure relates to the field of artificial intelligence, and in particular to a method, device and device for waking up an intelligent device. Background technique [0002] With the popularity of smart devices, wake-up technology has been widely used in life. For example, for smart devices such as smart phones and smart speakers, when people say corresponding keywords, the wake-up engine in the smart device will be activated, the smart device will be awakened from the sleep state to the working state, and then the smart device will be controlled by voice Corresponding operations, such as controlling smart devices to make calls, set alarm clocks, play music, etc., provide great convenience to life. [0003] The wake-up engine of the related technology is generally realized by a neural network model (referred to as the wake-up model), and in the wake-up scheme of the smart device of the related technology, after the smart device obtains t...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06F9/4401G06F40/30G06K9/62G10L15/22
CPCG06F9/4418G06F40/30G10L15/22G10L2015/223G06F18/241
Inventor 陈孝良许孝先常乐
Owner SOUNDAI TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap