Design And Development Of Language Rehabilitation Training System For Deaf-mute Children Based On Chinese Speech Analysis

Language is the most important form of communication between people. But in society, because of various reasons,there are so many children could not speak. The deaf children as a special group need great attention and support by society. Traditional speech rehabilitation training for deaf children is done by professional teachers in the professional institutions. But it’s of low efficiency by this way, and it cannot face a lot of students one time. Besides, the amount of professional teachers is small, so it makes deaf children’s language rehabilitation become a social problem.With the development of computer technology and language processing technology, it’s the trend of the times that making the speech rehabilitation training and computer technology combined. This article takes computer as the platform, designing and developing rehabilitation of deaf children’s language training system based on the analysis of Chinese phonetic. Using the digital speech processing technology and multimedia technology, this system presents the training contents in the form of human-computer interaction.This article starts with the mechanism of pronunciation, simulating people’s pronunciation process as the motivation model, the channel model and the radiation model. According to the characteristics of Chinese phonetics and the HSL theory, it designs the overall architecture of the system. This system is divided into three platforms, speech therapy, sound compose training and voice training, containing six modules, the pronunciation training, volume training, duration training, tone training, voiceless and voiced training and the words training. On base of researching the characteristics of Chinese phonetics, it puts forward the method of realizing each modules. With calculating the short-time average energy, it realizes the pronunciation training and volume training. With double judging the time threshold and the short-time average energy, it realizes the duration training. With the linear predictive coding technology(LPC) and autocorrelation function method, it realizes the tone training and sound and voicing decision training. With extracting the MFCC parameters of speech signal, using DTW(dynamic time warping) technology, it realizes the isolated word speech recognition. This system designs the corresponding training scenario for each training module, combining the multimedia image location and animated character movement with the speech processing date, and at last it presents the training contents in form of pictures and animations. Meanwhile, it evaluates the training result by human-computer interaction way, and inspires the interests of the trained children, then completes the deaf children’s language rehabilitation training preliminary.