Hi Prabhakar,

it depends on what kind of noise you have in your images. Is it additive or multiplicative (speckle noise)? If you can be more specific I will be able to give certain ideas.

Thank you for your response It is speckle noise i need despeckling methods

An efficient method for accurate noise filtration is based on the Fourier transform FT and Gaussian low pass filter GLPF. The quality of the de-noising process with FT and GLPF is strongly dependent on the value of the D0 parameter. To assess the performance of the de-noising process (to establish the optimum value of the D0 parameter), you can performed your test in two scenarios: when the additive Gaussian and the multiplicative Rayleigh noise exist in the images. Generally, you cannot find “noise-only” images, so you can generate noisy images by adding the noise into images, and then you calculate the optimal D0 value that minimizes the restoration error.

You can easily test FT convolution filters by using the free software package, ImageJ. http://rsb.info.nih.gov/ij/ ImageJ is written in Java and has a built-in fourier filter package which can be used to describe the filter you wish to use for removing your speckle noise. The documentation on the web site also includes this operation. A further discussion is carried out on the Wiki for this operation -- http://imagejdocu.tudor.lu/doku.php?id=gui:process:fft&s[]=fourier

Good luck with your studies.

Hi, good luck.

There is a tutorial page of mine,

http://www.academic.ro/ebook/profesor/html-cap6/cap6.htm

where ImageJA is implemented. In case of need,

We have both a seed of image processind program and a good programming environment, ProgrammingPool, tested for more than 10 years.

Dear Prabhakar,

I popeed into your question. Saw that you need despeckle algorithms. See us on resarch gate and visit our website http://www.medinfo.cs.ucy.ac.cy/

Total variation denoising, anisotropic diffusion are excellent performers. You can try the same.