Deaf Computer Intelligent Language Rehabilitation System

According to a survey of the disabled nationwide, there are about 20.57 million hearing handicaps, of which 800 thousand are under seven years old. The figure accelerates at the speed of 20-40 thousand every year. Some restoring training software for the deaf have been developed in the word .there are not these software for the Chinese deaf. So new restoration and training software is urgently needed to the China.So the aim of this research is to develop a kind of mutual visible language restoration training software by bimodal automatic speech. While analyzing the sound, it can provide visible means for language training and learning so as to make up for the hearing obstacle through vision.As the deaf s training vocabulary is rather small, most of time separate words are used ,so the principle of this system is to realize sound recognition by using DTW(Dynamic Time Warping) technology.The basic function is to provide two windows, one for training, the other for referencing, so that graphs and data of various analyzed results can be shown at the sane time on the screen. The former can be used as referenced standard sound, the latter is the pronunciation of the user. With the comparison, the problem of the user's pronunciation can be found exactly, so it can be helpful for their effective training and learning and correcting their wrong pronunciation so as to "Learn Speech Through Pictures", and greatly improve speech teaching and learning scientifically and directly. Meanwhile, the user's pronunciation can be automatically graded and recorded. In this way, the user can take an active part in the restoration training.