言语沟通障碍儿童辅助沟通及康复训练系统的研究与实现

发布时间：2018-05-12 05:46

本文选题：儿童言语沟通障碍 + 辅助沟通　；参考：《浙江大学》2017年硕士论文

【摘要】：我国存在言语沟通障碍的儿童数量较多,根据调查估计目前存在言语沟通障碍的儿童大约有一千万人,但国内针对这方面的研究、治疗相较于国外起步较晚,且目前国内对这些儿童的治疗方式以上特殊学校、医院,使用特殊设备进行一对一治疗为主。由于基数过大,这些手段很难惠及到所有的家庭,大量存在言语沟通障碍的儿童无法进行言语康复的训练,无法正常地和人交流。近年来,语音相关技术在各个领域开花结果,但是在语言沟通障碍儿童的辅助沟通、康复训练领域,基于常用移动设备如手机、平板电脑实现辅助沟通,在其中结合语音识别技术的相关成果、产品在国内并不多见,而且目前并没有引起足够的重视。本文结合国内外针对言语沟通障碍儿童的辅助沟通、康复训练的研究现状和发展趋势,结合语音相关技术的应用,对适用于存在言语沟通障碍的儿童言语沟通障碍康复训练系统在Android平台上进行了设计和实现。本文首先针对言语沟通障碍儿童的特点,设计了基于图标的辅助输入、沟通系统,用户只需在Android设备上通过轻敲相应的图标,就可以完成内容的输入,这些符号不仅自身都可以发音,输入完一个句子之后还可以自动生成自然的语音。在此基础上本文设计了针对单个词语和整个句子的言语康复训练功能,用户只需轻敲相应的词语、语句,系统便会生成相对应的语音,此时用户只需模仿系统的语音进行发音练习,系统便会自动生成对发音结果准确度的详细报告。本文同时对辅助沟通和康复训练两个功能进行了有机结合,康复训练的成果会直接影响辅助沟通功能的使用,两者相辅相成,形成激励机制。在此基础上,为了提高在使用辅助沟通系统的时候的输入效率,结合基于图标的输入系统的特点和针对言语沟通障碍儿童自身语言能力构建不完全的特征,在传统的基于N-gram的文本输入预测算法的基础上提出了将无语序的N-gram算法和基于语义的N-gram算法相结合的非句法性的混合算法,以提高实际输入时的效率。此外,为了让存在言语沟通障碍的儿童同样具有使用言语沟通的能力,本文在主流的语音识别技术的基础上,结合言语沟通障碍儿童自身的特点和言语训练子系统,设计出面向发音障碍人群的特定人语音识别系统,以适应这些儿童的发音。最后,本文在上述基础上设计并实现运行于Android平台上的言语障碍儿童辅助沟通及康复训练系统和起辅助性作用的网页端系统。
[Abstract]:There are more children with speech communication disorders in our country. According to the survey, there are about 10 million children with speech communication disorders. However, in China, the treatment started later than that in foreign countries. And the current treatment for these children above special schools, hospitals, using special equipment for one-to-one treatment. Because the base number is too large, it is difficult to benefit all families by these means. A large number of children with speech communication difficulties can not carry out speech rehabilitation training and can not communicate with people normally. In recent years, voice-related technologies have blossomed in various fields, but in the fields of language communication for children with communication difficulties, rehabilitation training, based on commonly used mobile devices such as mobile phones, tablets to achieve assisted communication, In combination with the related achievements of speech recognition technology, the products are rare in China, and have not been paid enough attention to at present. This paper combines the present situation and development trend of the research on auxiliary communication and rehabilitation training for children with speech communication disabilities at home and abroad, as well as the application of speech related technology. The rehabilitation training system for children with speech communication disorder is designed and implemented on Android platform. In this paper, according to the characteristics of children with speech communication disability, we design an aided input and communication system based on icon. The user can complete the input of content by tapping the corresponding icon on the Android device. These symbols can be pronounced not only by themselves, but also automatically by entering a sentence. On this basis, this paper designed a speech rehabilitation training function for a single word and a whole sentence. Users only need to tap the corresponding words and sentences, and the system will produce corresponding speech. In this case, users only need to imitate the pronunciation of the system for pronunciation exercises, the system will automatically generate a detailed report on the accuracy of pronunciation results. At the same time, this paper organically combines the two functions of auxiliary communication and rehabilitation training. The results of rehabilitation training will directly affect the use of auxiliary communication function. The two functions complement each other and form an incentive mechanism. On this basis, in order to improve the input efficiency when using the auxiliary communication system, combining the characteristics of the icon based input system and the language ability of the children with speech communication disabilities, we construct incomplete features. Based on the traditional text input prediction algorithm based on N-gram, a non-syntactic hybrid algorithm combining N-gram algorithm without word order and N-gram algorithm based on semantics is proposed to improve the efficiency of actual input. In addition, in order to make children with speech communication difficulties have the ability to use speech communication, this paper combines the characteristics of children with speech communication disorders and the speech training subsystem based on the mainstream speech recognition technology. Design specific speech recognition systems for people with dysphonia to accommodate the children's pronunciation. Finally, on the basis of the above mentioned above, this paper designs and implements the auxiliary communication and rehabilitation training system for children with speech disorders running on the Android platform and the web-side system which plays an auxiliary role.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R749.94;TN912.34

【相似文献】