用于语音识别中的SOFM矢量量化方法

该文讨论了神经网络语音识别系统中的ＫｏｈｏｎｅｎＳＯＦＭ神经网络的矢量量化（ＶＱ）的原理与过程,着重解决了其在实际应用时的若干问题,包括网络大小,学习步幅函数以及矢量量化过程中的公平竞争学习的控制函数的选定。过讨论了基于听觉模型的声学特性的提取与处理,这与ＶＱ性能息息相关。系统应用的实验表明,该神经网络在语音识别系统中具有良好性能。     

脑干听觉诱发电位波形分析建模、验证与算法优化

脑干听觉诱发电位是声觉器官受刺激后产生的一种生物电，反映了相应感觉通路及神经元群的活动，可用作听力阈值的客观判断，也可对多种耳疾和脑部病变作出诊断，因而成为当今临床医学和生物医学信号处理中的一个重要课题．文中介绍了脑干听觉诱发电位潜伏期（Ｌ）与刺激声强（Ｉ）之间关系曲线的作用，及其拟合建模和回归算法优化，还用实测数据对选定的拟合曲线模型进行了回归方差分析验证．     

A novel audio watermarking scheme using multiscale wavelet modulation

A novel audio watermarking scheme to embed robust and inaudible watermarks for the purpose of copyright protection is proposed. The key innovation is to add time-frequency redundancy into watermark signals by multiscale wavelet modulation. In order to maximize the watermarking strength within perceptual constraints, the signals synthesized from different scales are masked using a frequency auditory model, respectively, and then intergrated to form the final watermark signal. The detection structure is built using the redundancy in watermark signals, and the performance is further enhanced by modeling the statistical behaviors of wavelet coefficients as generalized Gaussian distribution. The use of original audio signal is not required in watermark detection. The experimental results show that our approach can achieve not only good transparency but also satisfying robustness to common audio manipulations.     

一种分层阈值优化的语音感知小波去噪方法

通过分析含噪语音信号的特点,引入能够兼顾人耳听觉特性的听觉感知小波变换,构造了新的小波阈值函数,并对小波变换分解后的阈值进行基于微粒群算法的分层优化.仿真实验表明,该方法在不同信噪比条件下均具有较好的去噪性能,语音的可懂度和听觉效果得到有效提高.     

基于听觉场景分析的近讲语音增强算法

针对近讲场景,提出一种双麦克近讲语音增强算法。该算法基于耳间延时差(ITD)、耳间强度差(IID)特征来区分目标语音,利用声学掩蔽效应,实现目标语音对背景噪声的分离降噪。与传统的单麦克增强方法相比,该算法可消除多种类型噪声且对语音造成的损伤较小。实验表明:该算法能将8～33dB的白噪声、音乐噪声、广播噪声3种噪声类型的带噪语音的信噪比提高到36dB以上,同时保持较高的目标语音可懂度。对于冲击噪声的带噪语音也具有较好的降噪效果。     

听觉线索引导大鼠空间定位

运用Morris水迷宫实验模式,研究听觉线索在大鼠进行空间定位时的作用.结果显示:听觉线索引导组和对照组动物在4 d内寻找平台的时间和错误率均显著下降.但与对照组动物相比,听觉线索引导组动物寻找隐藏平台的时间和错误率均较低,2组之间具有显著性差异(P<0.05).对动物的轨迹分析发现,听觉线索引导组和对照组动物在水池边缘区域分配的时间不同,提示2组动物寻找平台所用的策略不同.实验结果提示,大鼠能够利用单独的声音线索进行空间定位.     

A Novel Digital Audio Watermarking Scheme in the Wavelet Domain

We present a novel quantization-based digital audio watermarking scheme in wavelet domain. By quantizing a host audio‘s wavelet coefficients (Integer Lifting Wavelet Transform ) and utilizing the characteristics of human auditory system ( HAS), the gray image is embedded using our watermarking method. Experimental results show that the proposed watermarking scheme is inaudible and robust against various signal processing such as noising adding, lossy compression, low pass filtering, re-sampling, and re-quantifying.     

体温过低对大鼠脑干听觉诱发电位快慢成分的影响

目的：探讨体温过低对大鼠脑干听觉诱发电位快慢成分的影响。方法：计算机平均叠加技术颅表记录大鼠脑干听觉诱发电位（BAEP）快慢成分，体表物理降温法逐渐降低大鼠体温，传感探头式数字体温计监测大鼠直肠温度，体温每降低1℃测试1次BAEP快慢成分，观察波峰潜伏期（PL）、波峰间潜伏期（IPL）和波幅的变化。结果：随体温逐渐降低（36－25℃），BAEP快慢成分PL逐渐延长，快成分I－Ⅲ、Ⅱ-Ⅳ和I－Ⅳ IPL也逐渐延长，但两者波幅未随体温降低而发生规律性显著变化。结论：体温过低对大鼠BAEP的快慢成分均有显著影响。     

Topography of acoustic response characteristics in the midbrain inferior colliculus of Kunming mouse

Topography of acoustic response characteristics of the midbrain inferior colliculus(IC)of the Kunming mouse was studied by using extracellular recording techniques.The characteristic frequency(CF)range represented in the different divisions of the IC differed markedly:4-15 kHz in the dorsal cortex(DC),10-70 kHz in the central nucleus(CN),and 4-35kHz in the external cortex(EC).The CF in the CN increased from dorsal and lateral to ventral and medial,higher CFs represented at its ventromedial part and lower CFs at its dorsal part.The isofrequency contours of CFs were incurvate.Minimum thresholds(MT)of the auditory neurons in DC and the central part of CN were lower(about 10dB SPL),but considerably higher in the dorsal and ventral region of EC.Results suggest that each of the divisions in the mouse IC may have different auditory functions.     

黄雀、黄喉鹀延髓听觉神经核团的形态学研究与比较

用冰冻切片及尼氏染色法观察了鸣禽类黄雀(carduelisspinlaus)与黄喉鹀(Emberizaelegans)低位脑干中的听觉核团的位置、形态及细胞组成并进行了比较.结果表明,组成听觉上行通路第一级中继站的两对耳蜗核(Nucleuscochear),即延髓的巨细胞胲(Nuclcusmagnocellularis.NM)和角状核(Nucleusamgularis,NA)有明显的形态学差异.巨细胞核呈卵圆形,其细胞均为圆形或椭圆形,胞体较大,长径为25-30μm,突起较少.而角状核呈三角形,核内包含各种类型的细胞,其中有纺缍形、多极型和卵圆形等,胞体大小不等,长径为12-25μm,突起较长.由此提示这两对耳蜗核可能有不同的功能.延髓层状核(Nucleuslaminavis,NL)是听觉上行通路中的第二级中继站,核团呈棒状或卷曲状,中央部呈"Z”或"S”字形,细胞呈层状排列,胞体呈长梭形或多极型,长径为20-25μm,突起较长.本文为进一步研究鸣禽类的听觉机能及声源定位等提供了一定的形态学基础.