mcmc

对特征值进行非极大值抑制的角点检测，并且选用了合适的阈值加强检测效果(corner detection by non-maxima suppression of Eigenvalues and selected the appropriate threshold to enhance detection)

说明：  利用红黑小波进行图像去噪，图像复原，并与常规的小波进行比较(use red, black and wavelet denoising images, image recovery, and with the conventional comparison of wavelet)

稀疏贝叶斯学习的代码，出自UCLA，大家可以下载看看(sparse bayesian learning)

softcast的2d dct实现，先对图像转灰度图，得到的亮度信号矩阵做DCT 变换(softcast of 2d dct implementation, the first image to grayscale, luminance signal obtained by DCT transformation matrix do)

聚束SAR极坐标格式算法，包括dechirp处理以及距离向插值和方位向插值(Spotlight SAR polar format algorithm, including dechirp processing and distance and azimuth to the interpolation Interpolation)

编程实现大津（Otsu）自动灰度阈值分割算法的程序(Programming (Otsu) automatic gray Otsu threshold segmentation algorithm)

说明：  小波模极大源代码，可用来进行小波降噪或小波模态参数识别(Wavelet modal maximum source code, can be used for wavelet noise reduction or wavelet modal parameter identification)

。基于Wirtinger流的相位恢复算法是由Emmanuel J. Candes等于2015年提出的新的相位恢复算法，该方法不再将光场函数局限于非凸集中，而是利用一种新的初始化方式以及迭代规则来寻找最优解。(a new approach about phase retri which is published in 2015)

本程序是基于图论的知识进行图像分割，其中用到的是graphcut分割知识，完整的程序代码及结果！！(This procedure based on the knowledge of graph theory for image segmentation, which is used graphcut segmentation knowledge, including program code and the complete results !!)

计算局部形状描述符梯度方向直方图（HOG），直接输入图像即可用。(Calculate the local shape descriptors gradient orientation histogram (HOG), the input image can be used directly.)