Speaker recognition method, device, electronic device and storage medium
A speaker recognition and speaker technology, applied in speech analysis, instruments, etc., can solve the problems of weak feature expression ability and low accuracy rate of speaker recognition results, and achieve the effect of improving the accuracy rate
Active Publication Date: 2022-06-07
BEIJING CENTURY TAL EDUCATION TECH CO LTD
View PDF7 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, the feature expression ability extracted by using existing methods is weak, and the accuracy of speaker recognition results 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
[0085] In the following, only certain exemplary embodiments are briefly described. As those skilled in the art would realize, the described embodiments may be modified in various different ways, all without departing from the spirit or scope of the present application. Accordingly, the drawings and description are to be regarded as illustrative in nature and not restrictive.
[0086] figure 1 It is a flowchart of the speaker identification method according to the embodiment of the present application. like figure 1 As shown, the speaker recognition method may include:
[0087] Step S101, obtaining a target audio file and an audio file to be identified, and the target audio file includes the audio of the target speaker;
[0088] Step S102, dividing the target audio file and the audio file to be recognized into a plurality of audio units respectively;
[0089] Step S103, extract the corresponding audio feature from each audio unit, obtain the audio feature sequence of the t...
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 present application proposes a speaker recognition method, device, electronic equipment and storage medium. The specific implementation plan is: divide the target audio file and the audio file to be recognized into multiple audio units respectively; extract the corresponding audio feature from each audio unit, and obtain the audio feature sequence of the target audio file and the audio feature sequence of the audio file to be recognized ;Use the Siamese neural network to carry out feature learning on the audio feature sequence of the target audio file and the audio feature sequence of the audio file to be identified, and obtain the feature vector corresponding to the target audio file and the corresponding feature vectors of multiple audio units in the audio file to be identified ; Based on the feature vector corresponding to the target audio file and the corresponding feature vectors of a plurality of audio units in the audio file to be identified, the audio unit belonging to the target speaker in the audio file to be identified is identified using a machine learning model based on the attention mechanism. Using the embodiments of the present application can improve the accuracy of speaker recognition.
Description
technical field [0001] The present application relates to the technical field of speech recognition, and in particular, to a speaker recognition method, device, electronic device and storage medium. Background technique [0002] Speaker recognition is a technique to identify the speaker's identity through audio features. Existing speaker recognition methods roughly use the encoder model to encode the target audio file and the audio file to be recognized, and then judge the target detection result by comparing the encoded vector similarity. Most of the encoder models used are common. Deep neural network models, such as CNN (Convolutional Neural Networks, Convolutional Neural Networks) or RNN (Recurrent Neural Networks, Recurrent Neural Networks), etc. [0003] Taking the application scenario of classroom quality evaluation as an example, the number of times the teacher talks to the students and the interaction time between the two are considered as important indicators for e...
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): G10L17/18G10L17/02
CPCG10L17/18G10L17/02G10L17/08G06N3/096G06N3/0464G06N3/045G06N3/0442G10L15/02
Inventor 李航丁文彪刘子韬
Owner BEIJING CENTURY TAL EDUCATION TECH CO LTD