nonlocal-level-set

利用四参数随机生成法生成三维多孔介质，以研究黏土、页岩等的微观流动机制(Using QSGS to construct 3D porous media)

用自适应阈值法二值化图像，首先将图像分为几个子图像，在每个子图像中求得均值，用此均值二值化本子图(err)

立体视觉代码。主要过程包括：harris角点检测，zncc最优策略匹配，扩散得到更多特征对应，计算disparity和对应颜色值。另外提供了两幅图片以及对应的特征匹配的结果和disparity图像。(Three-dimensional visual codes. The main process includes: harris corner detection, zncc optimal strategy for matching, the proliferation of more features of the corresponding calculation of disparity and the corresponding color value. Also provides a picture and two corresponding feature matching results and disparity images.)

说明：  包括几种基于多项式插值的图像处理matlab代码，包括双三次插值，样条差值等(Includes several image processing based on polynomial interpolation matlab code, including double-cubic interpolation, spline difference, etc.)

自己编的加噪声的 很好用哦(Own noise increases by Oh well)

说明：  一种改进的均值聚类算法，能很好的利用与图像分割技术(k-means cluster)

该代码是基于学习的超分辨率重建，此算法是在线字典学习，字典的训练与稀疏表示与重建同时进行。(This code is based on the study of super-resolution reconstruction, this algorithm is learning online dictionaries, dictionaries of training and sparse representation and reconstruction simultaneously.)

傅里叶描述子，具有平移，旋转不变性，适合目标识别(Fourier descriptors, translation, rotation invariance, for target recognition )

Tracking based on background subtraction

图像处理-人脸识别技术：识别出图像中人物的五官形态，从而达到识别人脸的目的。(Image Processing-Face Recognition Technology)