speech-enhancement

说明：  ：该文首先研究了单认知用户能量检测算法的检测性能和检测灵敏度与检测时长和噪声短时间内平均功率波 动性之间的关系，得出检测性能和检测灵敏度随噪声短时间内平均功率波动性的增大而急剧下降的结论，在低信噪 比时尤为明显；针对检测性能和检测灵敏度随噪声短时间内平均功率波动性的增大而急剧下降，提出了基于多用户 协同能量检测算法。仿真表明，在低信噪比环境下，即使噪声短时间内的平均功率存在较大的波动性，只要适当地 增加协同检测的用户数，仍可获得较为准确的检测性能。(In this paper, detection performance and detection sensitivity with detection duration and noise average power fluctuation in short time is investigated firstly. Performance accuracy and detection sensitivity drop quickly with noise average power fluctuation increasing and it is worse when signal-to-noise ratio lower. Considering the characteristic, a novel new energy detection algorithm based on cooperative is presented. Simulations show that the proposed scheme improves antagonism of noise average power fluctuation in short time and gets a good detection performance as long as increasing the number of cooperative detection users.)

本压缩文件包含HHT-EMD(希尔伯特黄转换－经验模态分解)发明人黄锷院士领衔之台湾中央大学数据分析方法研究中心推广的标准EMD/EEMD/CEEMD分析工具，由该中心资深研究员Yung-Hung Wang王老师于2013年撰写，包含了原始EMD及近年来EMD的最新算法，最为权威、快速、精确、易用。 本人目前在该中心访学，对原文档进行了两处更动：1、加入程式注释中提到的5篇参考文献，以方便使用者随时参考，深入理解算法的意义；2、将eemd.m第111行代码getDefaultStream改为getGlobalStream(该行注释中已附参考链接)，解决了Matlab2013后的新版本在此运行出错的问题，旧版本用户可改回原代码。 纯干货，诚挚推荐。(fast EMD/EEMD/CEEMD code)

Face recognition for Yale face database

关于拉曼光纤激光器超连续谱的仿真程序，欢迎大家下载楼！(this is about supercontinue spectrum.)

自适应梯形求积法 特别适合被积函数为强峰的定积分 此程序修改余地很大(Adaptive trapezoid quadrature method was particularly suited to the plot function for the strong peaks of the definite integral of a large room for this procedure to amend)

对于具有关心的机械，采用2次函数进行预测，然后采用估计反求其合理值点(first we predict the possible velocity and then reverse the time and valuse by quadratic function)

Simulation program to realize OFDM transmission system(under one path fading)

GRNN网络设计，用交叉验证，循环的方法搜寻最优光滑因子。(GRNN network design, using cross-validation, the search for the optimal method of circulating smooth factor.)

一种基于蚁群算法的图像边缘检测算法，有中文注释。(Image edge detection algorithm based on ant colony algorithm)

极大化非高斯性的infomax ICA。和好很强大，希望对大家有所帮助(The maximization of non-Gaussian nature infomax the ICA. Very powerful and good, we hope to help)