The Design And Implementation Of Oral English Learning System Based On Speech Recognition Technology

Speech recognition is a technology that enables machine to transform speech signal into corresponding text or command through recognition and understanding process.In today's economic globalization,English has already become an important tool for people to communicate with each other.However,due to the constraints of learning patterns and environment,the level of spoken English is generally low.Based on the speech recognition technology,a spoken English learning system is designed and implemented.The user has more time,space advantages and diverse interactive characteristics when learning English.It not only helps to improve the learning effect,but also improves interest and enthusiasm in learning.The main contents of this paper include the following sections:First,it introduces the basic principle of speech recognition technology,explains the concept of Hidden Markov Model and three key algorithms,and realizes its simulation and implementation in the application of speech recognition,and introduces the architecture and key technology of the system.Second,the design and implementation of spoken English learning system.The main functions are as follows: firstly,the sample speech data is collected and preprocessed based on MATLAB,and the feature parameters extraction,codebook generation and template training based on the HMM are used to generate the fixed format codebook and template file.Secondly,MATLAB encapsulates the generated template files into a speech recognition library that can be called libspeech for the call of English spoken language learning system,and implements the transplantation of speech recognition technology on the Android platform.Finally,spoken English learning system is a APP application software developed on the Android platform based on libspeech speech recognition library.It mainly includes the design and implementation of user login,speech recognition,voice detection,voice broadcast,oral dialogue and other modules.Third,the recognition rate of different vocabulary size and different environment is tested and analyzed,which proves that the system has a good performance in performance.