pq_pcasvm

语音信号去噪，在MATLAB开发环境下，对含噪声的语音信号进行小波变换，实现语音增强的功能。(The speech signal denoising, wavelet transform, the noisy speech signal in the MATLAB development environment, Speech enhancement functions.)

希尔伯特变换和小波包结合的齿轮箱故障诊断方法，里面有具体的介绍小波包的分层方法，非常详细，可以下下来看看(Hilbert transform and wavelet packet of gearbox fault diagnosis method, there is detailed introduction inside the layered method of wavelet packet is very detailed, can down and have a look )

小波变换对电能质量暂态信号的分析和运用，包含电压骤升、骤降等(Wavelet transform for power quality analysis and the use of transient signals, including voltage dips, swells, etc.)

基于Wigner-Ville分布的Chirplet变换，用于信号的时频分析。

说明：  紧支撑小波基的构造。 以Daubechies小波滤波器系数为例，推导出小波函数及尺度函数，并画出其图像. void Convolution(double* pShortData, int nShort, double* pLongData, int nLong) int CalcScaleFunction(double* pOutData) int CalcWaveletFunction(double* pOutData) Convolution()不是通用的离散卷积代码，而是简单地实现了后面两个函数需要的离散卷积，它要求输入的前一个数组长度小于后一数组长度，而且卷积结果就直接覆盖存储在后一数组中。CalcScaleFunction()计算尺度函数，CalcWaveletFunction()计算小波函数。 (Compactly supported wavelet bases constructed. To Daubechies wavelet filter coefficients as an example, the derived wavelet function and scaling function, and draw their images. Void Convolution (double* pShortData, int nShort, double* pLongData, int nLong) int CalcScaleFunction (double* pOutData) int CalcWaveletFunction ( double* pOutData) Convolution () instead of generic discrete convolution code, but simply the realization of a function behind the need for two discrete convolution, which requires input before an array length is less than the array length of the latter, but the result of convolution directly covered by the latter is stored in the array. CalcScaleFunction () calculated scaling function, CalcWaveletFunction () calculation of wavelet function.)

用卷积深度置信网络CDBN的方法来处理睡眠EEG;(use Convolutional deep belief networks to sleep EEG)

说明：  近年来，随着互联网的高速发展，网上的多媒体信息也急剧增加，这些多媒体信息以图像为主。如何从浩瀚的图像数据库中快速、准确地找出所需要的图像，己成为一个备受关注的研究课题。有效地组织、管理和检索大规模的图像数据成为迫切需要解决的问题。于是基于内容的图像检索(Content-Based Image Retrieval: CBIR)作为一个崭新的研究领域出现了。 本课题拟研究、分析彩色图像红、绿、蓝三基色直方图的生成、特征提取和相似度等问题，应用图像的颜色信息—三基色直方图对图像进行检索。针对基于颜色的图像检索，本文采用应用广泛的RGB颜色空间来表示图像的颜色特征，对颜色分量进行等间隔量化并形成特征矢量并对特征矢量进行归一化处理，采用图像均匀分块的方法引入图像中色彩所处的位置信息，用距离度量函数进行图像的相似性匹配。在此基础上实现了基于三基色直方图算法的检索系统。 本文的研究和实践对于促进基于内容的图像数据库检索技术的应用具有一定的参考价值和实践意义。 (With the rapid development of Internet, the multimedia information is booming. All this information is mostly images. Effective recognizing, management and searching all these images have been an emergent problem. This has led the rise of a new research and development field: Content-Based Image Retrieval (CBIR). The topics to research, analysis color images red, green and blue color histogram generation, feature extraction and the similarity of the issues, application of the color image- trichromatic histogram of the image retrieval. Based on the color against the static image retrieval, this paper application of a wide range of RGB color space to indicate the color image features, the color components, such as spacing and quantitative characteristics of a feature vector and a normalization of vector processing, using uniform image block the introduction of the method in which the color image Location information, and distance measuring function similar to the image of the match. On)

小波变换(wavelet transform，WT)是一种新的变换分析方法，它继承和发展了短时傅立叶变换局部化的思想，同时又克服了窗口大小不随频率变化等缺点，能够提供一个随频率改变的"时间-频率"窗口，是进行信号时频分析和处理的理想工具。它的主要特点是通过变换能够充分突出问题某些方面的特征，能对时间(空间)频率的局部化分析，通过伸缩平移运算对信号(函数)逐步进行多尺度细化，最终达到高频处时间细分，低频处频率细分，能自动适应时频信号分析的要求，从而可聚焦到信号的任意细节，解决了Fourier变换的困难问题，成为继Fourier变换以来在科学方法上的重大突破。(Wavelet transform (WT) is a new transform analysis method. It inherits and develops the idea of localization of short-time Fourier transform and overcomes the shortcomings that window size does not vary with frequency. It can provide a frequency-dependent The Time-Frequency window is ideal for time-frequency analysis and processing of signals. Its main feature is that by transforming some features that can fully highlight some aspects of the problem, the localization of the time (space) frequency can be analyzed, and the signal (function) can be gradually and multi-scale refined by the scaling operation, finally reaching the high frequency Time subdivision, frequency subdivision at low frequencies, can automatically adapt to the requirements of time-frequency signal analysis, which can focus on any details of the signal to solve the difficulties of Fourier transform has become a major breakthrough in the scientific method since Fourier transform.)

2002年，MN Do和Martin Vetterli提出了一种“真正”的图像二维表示方法：Contourlet变换，也称塔型方向滤波器组(Pyramidal Directional Filter Bank, PDFB)。Contourlet变换是利用拉普拉斯塔形分解(LP)和方向滤波器组(DFB)实现的另一种多分辨的、局域的、方向的图像表示方法。(In 2002, MN Do and Martin Vetterli proposed a "real" two-dimensional representation of images: Contourlet transform, also known as Pyramidal Directional Filter Bank (PDFB). Contourlet transform is another multi-resolution, local and directional image representation method based on Laplacian pyramid decomposition (LP) and directional filter banks (DFB).)

这是三维离散小波变换程序，里面附带程序说明，及实例讲解。(This is a three-dimensional discrete wavelet transform, which comes with the program instructions and examples to explain.)