基于信号能量选取的电子耳蜗言语处理策略特征

Characteristics of cochlear implant speech processing strategies based on signal energy selection

基于信号能量选择的电子耳蜗言语处理策略的算法架构,分析该策略的信号处理过程,并比较波形、信号强度、语音频谱等特征,以获得该策略对语音信号处理的影响。此外,本文对信号序列每帧中选择的最大子带数量以及电子耳蜗滤波器组选择每个子带的概率进行深入统计。研究发现,基于信号能量选择的言语处理策略的特点是低频通道的选取概率较高,高频通道的选取概率较低。通过重点分析算法在信号特征参数传输上的匹配性和频带选择中的关键特征,为提出和改进新的电子耳蜗言语处理策略提供依据。

The study explores the algorithm architecture of cochlear implant speech processing strategies based on signal energy selection,analyzes the signal processing process,and compares several features such as waveform,signal intensity and speech spectrum for discussing the role of the strategy in speech signal processing.In addition,an in-depth statistical analysis is carried out on the maximum number of subbands selected in each frame of the signal sequence and the probability of each subband selected using the artificial cochlear filter bank.It's found that the speech processing strategies based on signal energy selection have the characteristics of high probability in low-frequency channels and low probability in high-frequency channels.The analysis on the matching of algorithms in signal feature parameter transmission and key features in frequency band selection provides a basis for putting forward and improving new speech processing strategies for cochlear implants.