Speech Recognition Technique In Oral English Learning

With the development of computer technology, computer-assisted technology in the field of education is widely used more and more. Now people can be more convenient to learn a language under the help of computer-assisted instruction. Rich computer graphics, voice processing function effectively promote people's language learning outcomes. The current research in this area is focused on the finding effective language learning methods, using the speech recognition technology combined with multimedia technology to develop teaching software with capacity of speech recognition and judgment. It has become a hot spot this type of language teaching.Although the current language teaching software develops fast, large number of software are in lack of evaluation and feedback of oral English pronunciation. However, in the language pronunciation learning, the effective feedback will do a great help to the learner. For this problem, this paper will study the Chinese people's English pronunciation habits, then establish a targeted corpus, and combined with requirement of Chinese-speaking learners , using HMM-based speech recognition technology to decode speech by Viterbi algorithm, and evaluate the pronunciation quality through a posteriori probability identification rate. Finally the system uses an expert library to do phoneme error correction, and achieve a system function integrated decoding, scoring and error correction. Main tasks are focused on in-depth study of speech assessment and error correction algorithms based on speech recognition, using speech recognition technology to solve a series of application difficulties in oral English learning, especially in the voice-in error correction. Ultimately provide a viable technical solution to the development of spoken English learning software.This paper describes the specific technical implementation of each module and provides detail of some core modules and core function. After system test, it uses testing to data demonstrate the feasibility of the system.This technology has been applied to the project "Word Expo Oral English Learning", which is developed by the Shanghai Jiao Tong University in cooperation with Sharp Corporation.