nlayer

图像小波变换，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。(Image Wavelet Transform)

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，根据阈值去噪的原理去噪，然后逆变换得到增强的语音信号(Speech Signal Denoising in the MATLAB development environment, including noise on the voice signal wavelet transform, in accordance with the principle of threshold denoising Denoising, and then inverse transform enhanced speech signal)

说明：  多媒体技术的发展和视觉信息的飞速膨胀迫切需要对视觉信息资源的有效管理和检索。由此，基于内容的图像和视频检索技术得到了越来越多的重视，成为了多媒体信息检索和图像处理领域中的重要研究方向。CBIR技术将对大规模图像信息的管理和访问提供有力支持。 本文这种介绍了内容图像检索的灰度特征实现方法，具有理论意义和实际应用价值。针对基于内容图像检索技术进行了研究，介绍了其研究现状和关键技术，讨论了其技术瓶颈和发展趋势。共生矩阵法，是对图像的所有像元进行统计调查，以便描述其灰度分布的一种方法。分析了这种基于广义图像灰度共生矩阵的图像检索方法，该方法首先将原图像作平滑处理得到平滑图像，然后将原图像和平滑图像组合得到广义图像灰度共生矩阵，提取该矩阵的统计特征量，最后将该统计量组成归一化向量用以检索。 实验结果表明，本方法的效果要优于单纯的灰度共生矩阵法。在对相似图像检索时，该算法与基于共生矩阵的经典算法相当，在面临同一图像大小显著改变或发生旋转的情况下，该算法明显优于经典算法。相比较而言，该算法提高了检索的有效性。 (The development of multimedia technology and the expansion of visual information，which cries for the effective management and retrieval methods of visual information. Thus, the image and video searching technology based on content is gaining more and more attention to become a multimedia information retrieval and image processing in the field of important research direction. The technology of CBIR provides effective supports for the management and access of extensive image information. This paper introduced the content of the image retrieval method characteristics of gray, with theoretical and practical value. This paper does an extensive and in-depth study of technology based on content image retrieval to, presenting their research status and key technologies, discussing its development trends and technological bottlenecks. Co-occurrence matrix of all the pixel image survey has been conducted in order to describe a method of distribution of gray. This paper analyses this broad-based )

离散小波变换的简单程序，适合初学者学习，附有通用函数。(Simple procedures discrete wavelet transform, suitable for beginners to learn, with common functions.)

将图像按位平面分解，然后在最低有效为嵌入一定的信息，达到数字隐藏的作用，是最原始的数字水印技术(image caput plane decomposition, and then embedded in the lowest effective for certain information, to hide the figures, is the most primitive digital watermarking technology)

说明：  利用小波变换将不同分辨率的图像进行小波分解，将高频和低频系数分别处理融合，进行小波重构。(Image fusion with different resolutions based on wavelet transform.)

说明：  直接减压到matlab\work目录下就可以看到运行结果了，其中包括了位平面分解，量化嵌入（提取），图像的金字塔分解等的程序块。(decompression directly to the Matlab \ work on the directory can see the results, which included a bit plane decomposition, quantification embedded (extraction), the image of the pyramid decomposition of the block.)

说明：   这是一个用小波神经网络进行非线性函数逼近的例子，拟合效果非常好，保证能运行！有详细注释！我是参考《青岛海洋大学学报》 2001年第1期 一种基于BP算法学习的小波神经网络编写的，大家可以针对原文更好的理解我的程序！(This is a wavelet neural network with nonlinear function approximation example of fitting results is very good, guaranteed to run! Detailed Notes! I refer to Journal of Ocean University of Qingdao，2001 ，“a learning algorithm based on BP wavelet neural network ”, you will understand me better for program!)

小波变换阈值去噪的文章，基于最大最小准则计算(Minimax Estimation, Adaptive Estimation, Nonparametric Regression, Density Estimation, Spatial Adaptation, Wavelet Orthonormal bases, BesovSpaces, Optimal Recovery.)

小波包的程序源代码,对于小波包的理解很有用，非常适合初学者！(Wavelet Packet program source code, useful for the understanding of wavelet packet, very suitable for beginners!)