基于频率调制信息的人工耳蜗语音处理算法研究  

Cochlear implant signal processing algorithm based on frequency modulation

作　　者：杨琳[1] 张建平[1] 王迪[1,2] 颜永红[1] 

机构地区：[1]中国科学院声学研究所中科信利语音实验室,北京100190 [2]中国科学院物理研究所,北京100190

出　　处：《声学学报》2009年第2期151-157,共7页Acta Acustica

基　　金：国家重点基础研究发展规划项目计划(973计划，2004CB318106);国家自然科学基金(10574140，60535030);国家高技术研究发展计划(863计划，2006AA010102，2006AA012195)资助项目。

摘　　要：在传统人工耳蜗连续交叠采样(Continuous Interleaved Sampler,CIS)算法的基础上,提出一种基于精细结构(频率调制信息)的人工耳蜗语音处理算法,在不引入过高频率成分、保证工艺可实现性的前提下,使语音识别率大幅提高。听觉仿真实验的结果表明,与传统的基于时域包络的CIS算法相比,基于精细结构的CIS算法对于元音可懂度的改进可以达到28%;声调的识别率在各种噪声条件下提高20%以上;在一般噪声环境下,辅音和句子的可懂度也分别获得了22.9%和28.3%的改进。A strategy was proposed to improve the performance of continuous interleaved sampling algorithm by introducing partial temporal fine structure cues, namely frequency modulation （FM） information, into the slowly varying temporal envelops. The improved algorithm has its own application values because it does not introduce too much high-frequency components into the model, which can not be perceived by deaf patients. The psychoacoustic experimental rtsults show that the introduction of the frequency modulation information can improve the cochlear implant （CI） performance greatly： the intelligibility of vowel perception can get over 28% improvements, tone recognition scores can increase more than 20% at various noise conditions, and under moderate noise conditions the consonant and sentence recognition can also be improved 22.9% and 28.3% respectively.

关 键 词：人工耳蜗 频率成分 语音处理 调制信息 算法 语音识别率 精细结构 噪声环境 

分 类 号：TP391.42[自动化与计算机技术—计算机应用技术]