File Exchange

image thumbnail

Detail Preserving Anosotropic Diffusion for Speckle Filtering (DPAD)

version 1.0 (9.36 KB) by

DPAD applies an anisotropic diffusion scheme to an image with multiplicative speckle



View License

DPAD implements two different anisitropic diffusion based schemes for speckle filtering:

-SRAD (Speckle Reducing AD)
-DPAD (detail Preserving AD)

Both methods are based on an anisotropic extension of Frost's and Kuan's LMMSE filters for multiplicative noise. The goodness of the filter is based on a proper estimation of the Coefficient of variation of noise.

The diffusion can also be done using an AOS operator that implements the semi-implicit scheme.

Methods in the paper explained in:

Yu Y, Acton ST. Speckle reducing anisotropic diffusion. IEEE Trans Image Process. 2002;11(11):1260-70.

S. Aja‐Fernandez, C. Alberola Lopez, On the Estimation of the Coefficient of Variation for Anisotropic Diffusion Speckle Filtering, IEEE Trans. Image Processing, Vol.15, no. 9, sept 2006.

Toolbox based on the SRADDIF toolbox implemented by Yu and Acton (Based on "Nonlinear Diffusion Toolbox" by Frederico D'AlmeidaFile ID: #3710)


Comments and Ratings (6)

laxman nainar



Jordan (view profile)


Derya (view profile)

Worked great for an image with uneven illumination - I was able to smooth out the background and increase the contrast in the image greatly. Great job, thanks for sharing. (though it's slow for large images)

What you have called I0 must be the input image. Note that you cannot name a variable with a number in matlab. Try Io instead. In addition, it is not an example, but an example of use. You must have some image Io and use it as your input.

Wu Jun

Wu Jun (view profile)

Please help me, the example can not run.

Example of use:
??? Undefined function or variable 'I0'.

MATLAB Release
MATLAB 7 (R14)

Inspired by: Nonlinear Diffusion Toolbox

Download apps, toolboxes, and other File Exchange content using Add-On Explorer in MATLAB.

» Watch video