Speech recognition method, device and electronic equipment
A technology for speech recognition and electronic equipment, applied in the computer field, can solve problems such as poor robustness of mixed models, difficulty in determining junk words, and difficulty in collecting audio samples of junk words, etc., to achieve the effect of improving accuracy
Active Publication Date: 2020-07-03
BEIJING ORION STAR TECH CO LTD
View PDF12 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
When training the above hybrid model, it is necessary to collect a large number of audio samples for keywords and audio samples for junk words. When the specific application scenario is determined, it is relatively easy to determine keywords, but it is really difficult to determine junk words, so it is even more difficult. It is difficult to collect audio samples for junk words, and because there are fewer junk words identified and fewer audio samples for junk words, the robustness of the hybrid model trained is poor, and then the hybrid model obtained by the above training is applied When performing speech recognition, the accuracy of the recognition result is low
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0050] The following will clearly and completely describe the technical solutions in the embodiments of the application with reference to the drawings in the embodiments of the application. Apparently, the described embodiments are only some of the embodiments of the application, not all of them. Based on the embodiments in this application, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of this application.
[0051] The concepts involved in the embodiments of this application are firstly introduced as follows:
[0052] 1. Keyword: This term is a relative concept, influenced by the device control commands that electronic devices can respond to. Specifically, it can be understood as a word that has a positive correlation with one or more device control commands that electronic devices can respond to.
[0053] For example, the electronic device is a smart speaker, and the device control commands th...
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
Login to View More Abstract
The embodiment of the invention discloses a voice recognition method and device, and electronic equipment and relates to the computer technology field. The method comprises the following steps of inputting the acquired audio to be recognized into a preset voice recognition model, carrying out voice recognition, and acquiring a first recognition result, wherein the voice recognition model is used for identifying whether there is the model of a key audio segment in an audio, and the key audio segment is the audio segment whose voice recognition result contains a preset keyword; under the condition that the audio to be recognized displayed by the first recognition result contains the key audio segment, inputting a target key audio segment to a confidence model so as to obtain a second recognition result, wherein the confidence model is used to obtain the confidence of the key audio segment, and the target key audio segment is the key audio segment contained in the audio to be recognized;and according to a second recognition result, determining whether the audio to be recognized is the audio used for awakening the electronic equipment. In the scheme provided in the invention, voice recognition is performed and the accuracy of a voice recognition result is increased.
Description
technical field [0001] The present application relates to the field of computer technology, in particular to a voice recognition method, device and electronic equipment. Background technique [0002] With the development of technology, voice has become the preferred way for users to interact with smart devices (such as: smart speakers, etc.). When the user interacts with the smart device through voice, the smart device must first perform voice recognition on the user's voice, and then control the device according to the recognition result to realize the interaction with the user. [0003] In the prior art, considering factors such as hardware performance and response speed, when smart devices perform speech recognition, speech recognition is generally performed through a pre-trained hybrid model based on HMM (Hidden Markov Model, Hidden Markov Model). When training the above hybrid model, it is necessary to collect a large number of audio samples for keywords and audio samp...
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
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G10L15/22
CPCG10L15/22G10L15/32G10L2015/223
Inventor 邹赛赛杨鹏贾磊
Owner BEIJING ORION STAR TECH CO LTD