本文为美国德克萨斯大学达拉斯分校（作者：KALYAN S. KASTURI）的电子工程硕士论文，共194页。
人工耳蜗是由植入电极和信号处理器组成的装置，设计用于恢复深度耳聋人群的部分听力。自上世纪70年代初人工耳蜗诞生以来，逐渐得到广泛普及，因此已经进行了大量的研究来推动并改进人工耳蜗技术。到目前为止，在人工耳蜗领域进行的大部分研究主要集中于改善安静无噪声（或低噪声）状态下的语音感知。在嘈杂的听力环境下，音乐和言语感知仍然是极具挑战性的问题。许多研究表明人工耳蜗在简单旋律识别任务中的识别分数较低。大多数人工耳蜗装置采用信号的包络幅度来提供电刺激。了解人工耳蜗植入过程中各种因素对旋律识别的影响，对于改进现有的编码策略具有重要意义。
本文研究了滤波器间距、相对相位、谱上移、载频和相位扰动等因素对声学听觉旋律识别的影响。目前用于人工耳蜗的滤波器间距大于音乐半音阶梯，因此并非所有音符都能够被解析。在当前的工作中，我们研究了称为“半音阶滤波器间距”的新技术，其中滤波器带宽根据音乐半音阶梯变化。迄今为止所研究的用于人工耳蜗的降噪方法大多是预处理方法。在这些方法中，首先使用降噪增强语音信号，然后使用语音处理器处理增强的信号。一种更好和更有效的方法是将降噪机制集成到人工耳蜗信号处理中。本文研究了两种嵌入设计的噪声抑制方法，即“信噪比加权法”和“S形压缩法”，分析了两种方法在噪声环境下改善语音感知的效果。信噪比加权降噪方法是一种指数加权法，使用瞬时信噪比(SNR)估计实现人工耳蜗中特定电极对应的每个频带的降噪。S形压缩法基于噪声估计将压缩曲线划分为两个区域，该方法对噪声部分和语音部分应用不同类型的压缩，因此与常规幂律压缩相比能够更好地抑制噪声。
Cochlear implants are prosthetic devices, consisting of implantedelectrodes and a signal processor and are designed to restore partial hearingto the profoundly deaf community. Since their inception in early 1970s cochlearimplants have gradually gained popularity and consequently considerableresearch has been done to advance and improve the cochlear implant technology.Most of the research conducted so far in the field of cochlear implants hasbeen primarily focused on improving speech perception in quiet. Musicperception and speech perception in noisy listening conditions with cochlearimplants are still highly challenging problems. Many research studies havereported low recognition scores in the task of simple melody recognition. Mostof the cochlear implant devices use envelope cues to provide electricstimulation. Understanding the effect of various factors on melody recognitionin the context of cochlear implants is important to improve the existing codingstrategies. In the present work we investigate the effect of various factorssuch as filter spacing, relative phase, spectral up-shifting, carrier frequencyand phase perturbation on melody recognition in acoustic hearing. The filterspacing currently used in the cochlear implants is larger than the musicalsemitone steps and hence not all musical notes can be resolved. In the currentwork we investigate the use of new filter spacing techniques called the‘Semitone filter spacing techniques’ in which filter bandwidths are varied in correspondenceto the musical semitone steps. Noise reduction methods investigated so far foruse with cochlear implants are mostly pre-processing methods. In these methods,the speech signal is first enhanced using the noise reduction method and theenhanced signal is then processed using the speech processor. A better and moreefficient approach is to integrate the noise reduction mechanism into thecochlear implant signal processing. In this dissertation we investigate the useof two such embedded noise reduction methods namely, the ‘SNR weighting method’and the ‘S-shaped compression’ to improve speech perception in noisy listeningconditions. The SNR weighting noise reduction method is an exponential weightingmethod that uses the instantaneous signal to noise ratio (SNR) estimate toperform noise reduction in each frequency band that corresponds to a particularelectrode in the cochlear implant. The S-shaped compression technique dividesthe compression curve into two regions based on the noise estimate. This methodapplies a different type of compression for the noise portion and the speechportion and hence better suppresses the noise compared to the regular power-lawcompression.
1 引言
2 人工耳蜗简介
3 文献回顾
4 利用人工耳蜗改善旋律识别的策略
5 噪声环境下人工耳蜗实现更优语音感知的策略
6 结论
下载英文原文地址：
/fs/flacbj02429132d9160/
更多精彩文章请关注微信号：
【信息技术】【.12】人工耳蜗在噪声环境中更好地识别旋律并改善语音理解的信号处理策略
如果觉得《【信息技术】【.12】人工耳蜗在噪声环境中更好地识别旋律并改善语音理解的信号处》对你有帮助，请点赞、收藏，并留下你的观点哦！
