Re: [OctDev] Greylevel morphological operations + other ideas

[Stefan v. d. W. <st...@su...>]

Hi Etienne

Adding grey-level morphology is a good idea.  Does MATLAB have any
similar functions?  What sort of interface do they use?

We should really also be looking at re-implementing octave-forge's
image handling.  Now that octave supports uint16 etc., we do not need
to guess image depths anymore.  Colour images should be stored as
MxNx3 matrices of the correct value type (not as three different images).

Unless someone else feels highly motivated, I will start on this
sometime closer to February (maybe by first adding some rudimentary
unit tests to all imaging functions, to track what breaks with the
changes).

Another feature I would like to add is the handling of videos.  I was
thinking of using libavifile -- does anyone here have any experience
with it?

Finally, the image histogram equalisation function is still not
working correctly.  But fixing it will be a lot easier once the image
types are implemented.

Regards
Stefan

On Thu, Jan 06, 2005 at 01:48:36PM -0500, Etienne Grossmann wrote:
> 
> 
>   Hi All,
> 
> I consider adding some greylevel image morphology functionality to
> octave-forge. I have this function:
> 
> 
> ## im2 = g_all (im,scales)       - Greylevel morphological opertations
> ##
> ##      Open, close, erode and dilate greylevel image im at various scales.
> ##      Dilation at scale s is the greatest greylevel within radius s.
> ##      Erosion  at scale s is the least greylevel within radius s.
> ##      Opening  is erosion followed by dilation (all at the same scale).
> ##      Closing  is dilation followed by erosion (all at the same scale).
> ## 
> ## im     : R x C       : Image
> ## scales : 1 x S       : Scales at which morphological operations are performed
> ##
> ## im2     : R x C x 4*S : Opened (4*i+1), closed (4*i+2), eroded (4*i+3) and
> ##                        dilated (4*i+4) images (i in 0:S-1)
> ##
> ## See also: g_open_close.
> ## Todo    : Make this function more efficient.
> 
>   Any suggestions on other possible interfaces? I can also put
> fucs for individual operation.
> 
>    Cheers,
> 
>   Etienne
> 
> -- 
> Etienne Grossmann ------ http://www.cs.uky.edu/~etienne
> 
> 
> -------------------------------------------------------
> The SF.Net email is sponsored by: Beat the post-holiday blues
> Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
> It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
> _______________________________________________
> Octave-dev mailing list
> Oct...@li...
> https://lists.sourceforge.net/lists/listinfo/octave-dev