摘要
目的利用非局部平均(NL-means)算法提高高刺激率模式下去卷积前的听觉诱发电位(AEPs)的信噪比。方法首先通过仿真实验来确定在不同噪声水平下的滤波参数,然后采用真实数据进行了验证,并且与均值滤波进行了对比。结果 NL-means方法能够有效的抑制粉红噪声,较好的保留信号的细节。结论 NL-means方法非常适合具有周期特性的信号去噪,能够有效的去除噪声,提高AEPs信噪比。
Objective To enhance the signal to noise ratio (SNR) of the transient auditory evoked potentials (AEPs) with the high rate stimulation paradigm by the non-local means (NL-means) algorithm. Methods Ex- periments were conducted on simulation data to estimate the filtering parameters and AEPs from human sub- jects to demonstrate the validity of the proposed theory. The NL-means algorithm and the mean filtering were also compared. Results The NL-means algorithm could not only eliminate the pink noise effectively, but also preserve the details of the signal. Conclusion The NL-means algorithm is suitable for denoising periodic signal and can enhance the SNR of AEPs effectively.
出处
《航天医学与医学工程》
CAS
CSCD
北大核心
2014年第1期26-31,共6页
Space Medicine & Medical Engineering
基金
国家自然学基金资助项目(61172033
61271154)
关键词
听觉诱发电位
非局部平均算法
均值滤波
auditory evoked potentials
non-local means algorithm
mean filtering