Multi-model adaptive and speech recognition device and method

Abstract

The present invention includes: selecting any one model designated by a speaker; extracting a feature vector used in a voice model from an inputted voice of the speaker; adapting the extracted feature vector by using a predetermined pronunciation information model and a predetermined basic voice model and thereafter, storing the corresponding feature vector in a model designated by the speaker among the plurality of models, and setting a flag indicating whether adaptation is executed; extracting a feature vector from a voice which the speaker inputs for voice recognition; selecting only models in which adaptation is executed by reading flags set in multi adaptive models; calculating similarity of adaptive values by sequentially comparing the models selected by reading the flags with the feature vectors extracted from the inputted voices of the speakers; and selecting one model having the maximum similarity and executing voice recognition through decoding when similarity calculation for all the selected models is completed.

Description

technical field [0001] The present invention relates to a speech recognition system, more particularly to a multi-model adaptive and speech recognition device and its method, whereby the speaker's speech according to the characteristics is adapted and stored in the basic speech model and an independent new speech model. It provides stable real-time speech recognition through speech recognition using multivariate adaptive models. Background technique [0002] A voice recognition system does not have a model for each user, but seeks to recognize a person-independent voice with a single unspecified model so that the voices of an unspecified majority of people can be recognized. [0003] Since speech recognition is performed by statistical modeling, which is the underlying technology, the recognition rate varies depending on the speaker and also varies depending on the surrounding environment. [0004] The problem of lower recognition rate due to the surrounding environment can...

AI Technical Summary

Problems solved by technology

[0013] Therefore, when multiple people with unsatisfactory speech recognition sequentially adapt the speech recognition system, the recognition rate will increase sharply at first, but as the adaptation is performed sequentially, the recognition rate of the first adaptive speaker will gradually decrease, and only the last The recognition rate of the adaptive speaker is relatively good

Images