*国家自然科学基金项目资助 (60970136); 中国科学院声学研究所所长择优基金资助 (Y154221341)
Δ 通讯作者。 Email: hxlhan123@163.com 1
扫频激励下的 DPOAE 的局部加权最小二乘估计算法
*
韩晓丽
1Δ
傅新星
2,3
崔杰
1
肖灵
1
1 (中国科学院 声学研究所,北京 100190)
2 (首都医科大学附属北京同仁医院,北京 100730 )
3 (北京市耳鼻咽喉科研究所,北京 100730 )
摘要:畸变产物耳声发射(distortion product otoacoustic emissions, DPOAE)信号可以用于听力
损失的诊断,具有重要的临床应用价值。但是,大频率范围内进行DPOAE测量时,需要耗
费大量的时间。采用扫频信号作为诱发信号,为记录大频率范围内的DPOAE测量数据提供
了一种快速而有效的手段。本文在已有的最小二乘拟合(least-squares-fit, LSF)算法的基础上,
给出一种估计扫频激励下的DPOAE信号的局部加权最小二乘估计算法(locally weighted least
squares estimation, LWLSE),从扫频激励下的DPOAE记录信号中提取方差根音(cubic different
tone, CDT)频率处的DPOAE信号。仿真结果表明,该方法相对于常用的LSF方法,对DPOAE
信号和初始刺激信号的估计更准确,并且估计结果具有更小的估计方差,有利于提取更加准
确的DPOAE精细结构。
关键词:畸变产物耳声发射 最小二乘拟合 加权最小二乘估计 精细结构
中图分类号:TN911.7
文献标识码:A
Locally Weighted Least Squares Estimation of DPOAE Evoked by Continuously Sweeping
Primaries
Han Xiaoli
1
Fu Xinxing
2,3
Cui Jie
1
Xiao Ling
1
1 ( Institute of Acoustics, Chinese Academy of Sciences, Beijing 100190, China)
2 (Beijing Tongren Hospital, Capital Medical University, Beijing 100730, China)
3 ( Beijing Institute of Otorhinolaryngology, Beijing 100730, China )
Abstract: Distortion product otoacoustic emission (DPOAE) signal can be used for diagnosis of
hearing loss and it has important clinical value. Using continuously sweeping primaries to
measure DPOAE provides an efficient tool to rapidly record DPOAE data when DPOAE is
measured in a large frequency range. In this paper, locally weighted least squares estimation
(LWLSE) of cubic different tone (CDT) DPOAE is presented based on least-squares-fit algorithm,
in which DPOAE is evoked by continuously sweeping tones. A weighted error function is used as
the loss function and the weighting matrixes are in the local sense to obtain a smaller estimate
variance. First, ordinary least squares estimation of the DPOAE parameters is obtained. Then the
error vectors are grouped and the different local weighting matrixes are caculated in each group.
At last, the parameters of the DPOAE signal are estimated based on least squares estimation
principle using the local weighting matrixes. The simulation results show that estimate variance
and fluctuation errors are reduced, thus the method estimates DPOAE and stimuli more accurately
and stably, which facilitates extraction of clearer DPOAE fine structure.
Key words: DPOAE LSF WLSE Fine structure
1. 引言
当两个频率相近的纯音信号(
1
,
2
;
f
21
> )同时刺激耳蜗时,可以在外耳道中检测到畸
变产物耳声发射(distortion product otoacoustic emissions, DPOAE)信号。DPOAE的产生是耳蜗
非线性特性和两个初始刺激音调互调制作用的结果。研究发现,DPOAE信号至少包含两个
主要构成成分——畸变产生成分和反射成分。这些不同的构成成分共同组成了在耳道中记录