A method and device for improving the accuracy of speech recognition

Abstract

An embodiment of the invention discloses a method and device for improving speech recognition accuracy. The method comprises the following steps: obtaining speed information from a user terminal, andaccording to the telephone number of the user terminal, obtaining family network information of a user; according to the family network information of the user, obtaining the telephone number of eachfamily user meeting preset conditions, and obtaining home location information of the telephone number of each family user; according to the home location information of the telephone number of each family user, obtaining a target home location, and determining a target acoustic model and a language model according to the target home location; and according to the target acoustic model and the language model, converting the speech information into text information. The device is used for realizing the method. The method for improving speech recognition accuracy can improve application accuracyof the acoustic model and the language model, and furthermore, improves speech recognition accuracy.

Description

technical field [0001] The invention belongs to the technical field of communication, and in particular relates to a method and a device for improving the accuracy of speech recognition. Background technique [0002] With the development of speech recognition technology, applications based on speech recognition technology emerge in endlessly. Especially in the communication field, major communication operators have launched applications and services based on speech recognition technology. The accuracy of speech recognition is directly related to the success of the application or business. [0003] Generally, there are several ways to improve the accuracy of speech recognition: [0004] 1. Optimizing the acoustic model: This method obtains the optimal correspondence between speech and syllables by calculating the probability of speech to syllables. At present, the acoustic model is mainly based on the Hidden Markov Model (HMM) framework for acoustic modeling. The maturity...

AI Technical Summary

Problems solved by technology

[0009] It is not possible to customize and optimize the user's speech system: Since the speech expressions of users in each region will have different local characteristics, the online acoustic model and language model are adaptively optimized based on the online data, although the overall optimization scheme The versatility is relatively good, but for user groups with obvious regional differences or unique pronunciation, it is difficult to maintain a relatively accurate recognition effect

Even though different acoustic and language models are established for user groups in various regions and accents in the existing technology, in actual application, it is not possible to quickly and accurately confirm the region or accent category of the user, making the acoustic and language models applicable incorrectly.

Images