DenoisingMicrograph

micrograph_denoise

Purpose

This utility allows you to apply denoising algorithms to micrographs. The output image is a raw file with floats (32 bits). The most important denoising method for micrographs is the Mumford-Shah difussion, this method is based on the paper:

Teboul, et al. IEEE-Trans. on Image Proc. Vol. 7, 387-397. 

Usage

$ micrograph_denoise -i [image in] -o [image out]

The output image must be different from the input one since both images are dealt at the same time from disk. The input raw file must be accompanied by the corresponding.inf file (see__xmipp_mark__ for a description of this file).

Parameters

• `` Input file
• `` Output file
• -window_size [n128] = Size of the window that is treated independently. This parameter limits the amount of memory required.

For the denosing specification you can supply any of the denoising parameters.

Examples and notes

• The Mumford-Shah difussion is based on the derivative of the image. On the image limits this derivative cannot be computed so accurately, so it's normal a frame of "noisy" pixels whose size is equal to the number of outer iterations.

Denoise a micrograph with the default parameters using Mumford-Shah difussion:

$ micrograph_denoise -i piece.raw -o filtered_piece.raw -denoising difussion

The results are shown in the following pictures:

Original	Filtered
/piece.gif	/filtered_piece.gif

--Main.AlfredoSolano - 16 Jan 2007