[sane-devel] ANNOUNCEMENT: dust removal: please test

[Peter Seiderer]

Hello Andras,

did some test with your 'duster' and with a very similar own algorithm, see
 http://www.ciselant.de/projects/antidust/

Your duster seems to miss some pixels and/or areas of dust.
Maybe test if you hit enough random choosen pixel to calcuate the
interpolation and iterate some times over the nx,ny loops (rembering the
already interpolated pixels).

 Peter

On Tue, Aug 20, 2002 at 10:59:50PM +0100, Major A wrote:
> 
> Hi all,
> 
> I'm proud to make available the first testing version of a
> dust-removal facility for film scanners with IR readout. Currently it
> is a stand-alone tool, but the actual algorithm is implemented in a
> sanei library. The algorithm is so simple I can't think of a good
> reason why I didn't come up with it earlier.
> 
> To get it, check out the latest CVS of the coolscan2 project on
> sourceforge.net (http://coolscan2.sourceforge.net/). Then use the
> program "duster" in the following way:
> 
>   duster <rgb-image> <ir-image> <out-image>
> 
> The input images are the RGB and IR data obtained from a scan with the
> coolscan2 driver. They can be in any format, in theory, but I've only
> tested PNMs. The output file is written in PNM 8-bit format with a
> gamma of 2.0 automatically applied (I used raw scans with neutral LUTs
> for testing, that's why).
> 
> The thresholding algorithm is really really simple, expect some
> improvement there later. It seems to work with my E6 scans, but it is
> not very robust.
> 
> duster is only temporary, the functionality will eventually all be
> included in some part of the SANE backend-frontend chain.
> 
> Please try this code and let me know what you think.
> 
>   Andras
> 
> ===========================================================================
> Major Andras
>     e-mail: andras at users.sourceforge.net
>     www:    http://andras.webhop.org/
> ===========================================================================

-- 
------------------------------------------------------------------------
Peter Seiderer                     E-Mail:  Peter.Seiderer at ciselant.de