Accent recognition model building method, device, storage medium and electronic equipment

Abstract

Embodiments of the present invention provide an accent recognition model establishment method, device, storage medium and electronic equipment. The method includes: respectively adopting the user voice training sample set of each first city to train the single-city accent recognition model of the corresponding first city to obtain the single-city accent recognition model of each first city; The user speech test sample set of the city is input to the single-city accent recognition model of each first city; for each second city, select among all the single-city accent recognition models of the first city: the recognition rate is the highest and the recognition rate is greater than the first city A threshold model, the first city and the second city corresponding to the selected model are divided into the same accent area; for each accent area, the user voice training sample set of all cities in the accent area is used The regional accent recognition model of the accent region is trained to obtain the regional accent recognition model of the accent region. The embodiments of the present invention improve the recognition rate of accent recognition.

Description

technical field [0001] Embodiments of the present invention relate to a method, an apparatus, a readable storage medium, and an electronic device for establishing an accent recognition model. Background technique [0002] At present, in housing transactions, for convenience, online VR (Virtual Reality, virtual reality) technology has been applied to show the house for customers. In this scenario, the agent needs to show the client the property as the VR camera changes. During the viewing process, the agent and the client will communicate, that is, both the agent and the client will speak. Since clients may come from different regions, there may be situations where brokers cannot fully understand what clients are saying. [0003] At present, there are some models that recognize speech as text. Input the user's speech in any region into the model to get the text corresponding to the speech. However, due to the large difference in accents between some regions, the recognition ...

AI Technical Summary

Problems solved by technology

Since customers may come from different regions, there may be situations where the broker cannot fully understand the customer's speech

[0003] At present, there are some models that recognize speech into text. Input the voice of users in any region into the model to get the text corresponding to the speech. However, due to the large differences in accents between some regions, the recognition rate cannot be guaranteed.

Images