Multilingual Speech Recognition Method Based on Collaborative Classification of Language Type and Speech Content

Abstract

The invention discloses a multilingual speech recognition method based on the cooperative classification of language types and speech content. The method includes: step 1) establishing and training an acoustic model for the cooperative classification of language types and speech content; the acoustic model incorporates language-related information Language feature vector, in the process of multilingual recognition, the language feature vector can be used to perform model adaptive optimization on the phoneme classification layer of a specific language; step 2) input the speech feature sequence to be recognized into the trained language type and speech content collaborative classification acoustic model, output the phoneme posterior probability distribution corresponding to the feature sequence; the decoder combines the sequence phoneme posterior probability distribution of the feature to generate several candidate word sequences and their corresponding acoustic model scores; step 3) combine the acoustic model scores of the candidate word sequences and The language model scores are combined as the overall score, and the candidate word sequence with the highest overall score is used as the recognition result of the speech content of the specific language.

Description

technical field [0001] The invention relates to the field of multilingual speech recognition, in particular to a multilingual speech recognition method based on cooperative classification of language types and speech content. Background technique [0002] At present, the automatic speech recognition technology is very mature. Under the technical research of some speech recognition institutions, the recognition accuracy of the automatic speech recognition system can reach 94.5%, which can be said to have reached the human auditory perception ability. But this excellent automatic speech recognition system is limited to a few widely used languages, such as English and French. There are more than 5,000 languages ​​spoken by people all over the world, but only 10 of these 5,000 languages ​​are widely used, they are: Chinese, English, Russian, Spanish, Hindi, Arabic , Portuguese, Bengali, German and Japanese. For other languages, due to the small number of users, it is difficult...

AI Technical Summary

Problems solved by technology

This method requires prior knowledge of language-related information in the model training and testing stages, requires a lot of manual annotation work, and the performance of speech content recognition is not high

Images