Oral Pronunciation Discrimination in English Education Based on Multimedia Data and HMM

In order to provide students with sufficient oral teaching resources and oral training platform to effectively improve students' comprehensive English application ability, a pronunciation discrimination and speech analysis system for English education based on multimedia data was proposed. The system uses the recorded standard pronunciation to train the model, which is used to score the phonemes of the pronunciation, build an expert knowledge base for common pronunciation errors, and provide corrective suggestions for the wrong pronunciation of the user. The experimental results show that the recognition speed is 10 times higher than that of PC, and the recognition accuracy is more than 70%. Conclusion: The pronunciation discrimination and speech analysis system based on multimedia data is of great practical significance and can provide a better oral teaching resources and training platform for students.