默声音符编码文本输入方法  

A text input method based on silent voice symbol encoding

作　　者：滕鹏[1] 贾云得[1] 

机构地区：[1]北京理工大学计算机学院,智能信息技术北京市重点实验室,北京100081

出　　处：《科学通报》2011年第4期354-359,共6页Chinese Science Bulletin

基　　金：国家自然科学基金资助项目(90920009,60905006)

摘　　要：人们希望语音输入能够取代键盘,成为计算机文本输入的最主要方式.但语音输入在噪声环境中准确率低,且会泄露输入信息,也会影响他人.本文提出默声音符编码文本输入方法,该方法选择默声音中容易发声、容易精准识别且彼此间容易区分的少数几个音素作为默声音符(silent voice symbol),对各种字符进行编码.类似摩尔斯电码,用户只需要记住编码方式,就可以通过读默声音符准确地实现文本输入.实验证明提出的方法有潜力替代键盘在各种环境下完成文本输入任务.Speech input is expected to become the principal method seriously degraded in nosiy environment, and often makes of computer text input and replace the keyboard. But speech input is a leakage of information as well. To avoid these problems, this paper presents a text input method based on silent voice symbol encoding. We select several easy-to-pronounced and easy-to-recognized phonemes in silent voice as silent voice symbols （SVSs） to encode characters. Similar to Morse code, memorizing the encoding scheme, users can input text accurately by speaking relevant SVSs. Experimental results demonstrate that the proposed method is promising to be a reliable method of text input for many applications.

关 键 词：文本输入 默声音符 无声交流 无声语音接口 

分 类 号：TP391.14[自动化与计算机技术—计算机应用技术]