中文说明：
包括图像分析的四部分代码：matlab扩散和高斯函数，线性扩散，线性复扩散，非线性扩散。
包括四个部分：
[1]函数：diffusion.m gauss.m
[2] 线性扩散将线性扩散应用于图像，创建线性比例空间。MATLAB代码：demo_lin.m Image:haifa1.bmp
[3]线性复扩散应用线性复扩散创建高斯和拉普拉斯尺度空间。MATLAB代码：demo\cmplin.m
[4]非线性扩散：Perona-Malik
[“各向异性扩散”]Catte等.P-m复斜坡保持扩散的正则化非线性边缘保持扩散经典Perona-Malik过程：扩散系数的值在一阶导数估计的边缘附近减小。最好应用于台阶边缘。Catte等人的正则化版本-用于控制过程的梯度估计通过高斯平滑。被证明在数学上是适定的。斜坡保持复杂扩散-最好的斜坡类型的边缘。结果更平滑，几乎没有楼梯效果。MATLAB代码：demo\nldif.m图片：ct_扫描.bmp
English Description：
Including four parts of image analysis code: MATLAB diffusion and Gaussian function, linear diffusion, linear complex diffusion, nonlinear diffusion. < / P > < p > includes four parts: < / P > < p > [1] function: diffusion. M Gauss. M < / P > < p > [2] apply linear diffusion to image to create linear scale space. Matlab code: Demo_ lin.m Image:haifa1.bmp < / P > < p > [3] linear complex diffusion uses linear complex diffusion to create Gaussian and Laplacian scale spaces. Matlab code: demo / cmplin. M < / P > < p > [4] nonlinear diffusion: Perona Malik < / P > < p > ["anisotropic diffusion"] catte et al. P-M complex slope preserving diffusion regularized nonlinear edge preserving diffusion classical Perona Malik process: the value of diffusion coefficient decreases near the edge of the first derivative estimation. It is best applied to the edge of the step. Catte et al's regularized version - gradient estimation for control processes via Gaussian smoothing. It is proved to be well posed mathematically. Slopes keep complex diffusion - the best slope type at the edge. The result is smoother, with little stair effect. Matlab code: demo / nldif. M picture: CT_ Scan.bmp