Self-adapting method aiming at computer language learning system pronunciation evaluation

Abstract

The invention relates to a self-adaptive method for evaluating the pronunciation of a computer language learning system, comprising the following steps of: establishing a voice recognition system; carrying out the segmentation of initials and finals of voice data based on the voice recognition system and restricting the recognition of initials and finals at the boundary and the tone; calculating posterior probability of all the initials and finals and the tone according to the result of restriction of boundary recognition and the segmentation and selecting self-adaptive data according to the preset threshold; carrying out self-adaptation for an acoustic model in the recognition system according to the selected self-adaptive data; carrying out the second segmentation and recognition by using the acoustic model after the self-adaptation; using the final segmentation and recognition result to extract evaluating parameters; the invention selects a proper self-adaptive corpus by posterior probability, not only can reduce error of a recognizer caused by the differences of an actual speaker and a standard model tone, use environment and channels but also can avoid the bias of the standardmodel when in self-adaptation so as to be incapable of evaluating the pronunciation of the speaker accurately.

Description

Technical field [0001] The invention relates to a method for making adaptive adjustments for pronunciation evaluation of a computer language learning system. Background technique [0002] In order to make it more convenient and more accurate for students to learn various languages, at present, more computer-aided language learning systems are used, but the current computer-aided language learning systems will have a problem: the number of speakers using standard acoustic model training data The scope is limited and it is impossible to cover various timbres; the standard corpus recording environment (office environment) is different from the actual environment (exam and learning environment), and the recording equipment may also be quite different. Affected by these factors, there is a large mismatch between the acoustic model and the actual use situation; these factors cause a large difference between the candidate’s pronunciation vector and the standard acoustic model in practic...

AI Technical Summary

Problems solved by technology

[0003] Aiming at the problems in the existing computer-aided language learning system, the present invention selects suitable adaptive corpus through the posterior probability to ensure that the recognizer error caused by the difference between the actual speaker and the standard model timbre, use environment and channel can be weakened, It can also prevent the standard model from being biased during self-adaptation, and cannot correctly evaluate the pronunciation of the speaker

Images