Description: It contains four parts：
[1]MATLAB function:diffusion.m gauss.m
[2]Linear diffusion
Applying linear diffusion to images creating linear scale-space.
MATLAB code: demo_lin.m
Image: haifa1.bmp
[3]Linear complex diffusion
Applying linear complex diffusion creating Gaussian and Laplacian scale-spaces. MATLAB code: demo_cmplin.m
[4]Nonlinear diffusions:
Perona-Malik ["Anisotropic diffusion"]
Catte et al. regularization of P-M
Complex ramp-preserving diffusion
Nonlinear edge preserving diffusions.The classical Perona-Malik process: the value of the diffusion coefficient is reduced near edges estimated by the first derivative. Best applied to step edges. A regularized version by Catte et al- the gradient estimation for controlling the process is smoothed by a Gaussian. Proved to be mathematically well-posed. Ramp preserving complex diffusion- best for ramp-type edges. Results are smoother with almost no staircasing effects.
MATLAB code: demo_nldif.m
Image: ct_scan.bmp
File list (Check if you may need any files):
图像程序代码
............\demo_cmplin.m
............\demo_lin.m
............\demo_nldif.m
............\diffusion.m
............\gauss.m
Download