Login  Register

Re: Convert a set of points into an ROI

Posted by karo03 on May 12, 2009; 6:09am
URL: http://imagej.273.s1.nabble.com/Convert-a-set-of-points-into-an-ROI-tp3692520p3692525.html

Hi Ghislaine,
I don't know about the problem of the origin being hen or egg, however,
if the hen is the set of isolated cells in an image with Analyze-
 >Analyze Particles "Add to Manager" activated you will get the roi of  
each cell (for any purpose),
if the egg is the set of points you can generate an image with one  
object set of points drawn, apply perhaps a small closing, "make  
binary" and edit->selection->create selection and add to manager to  
have a roi of that point set in the roi manager.

Hope I have not misunderstood you
Regards
Karsten

Am 12.05.2009 um 01:09 schrieb Ghislain Bonamy:

> Curtis,
>
> Thanks for the reply. Unfortunatelly, most of the forms are not  
> convex, but
> all of the points lie on the Hull. I have attached a zip file,
> http://n2.nabble.com/file/n2865999/ROI.zip ROI.zip , that contains: a
> picture showing the different ROI (different colors=different  
> objects), and
> the coordinates for thes objects, which provide the outline of these
> objects. The x;y vaues are separted by ';' and the different points  
> are
> separated by ','. This is not the final schema for the ROI format,  
> but it
> should give you an idea about what I have cooking.
>
> So the problem at hand is to be able to either order the points for  
> each
> objects in an order that is suitable for the registration of the  
> ROI. Could
> anyone explain what are the different type of ROI? In particular  
> what is
> polyline? This might be the best type to use, but I do not know what  
> it
> represents.
>
> I think that one garante with the ROI outputed, is that for each  
> pixels
> provided, there are only 2 adjacent points (ie, if you imagin a  
> pixel, it
> can only be surounded by 8 pixels, since we are in a 2D space). So one
> strategy, would be to take one point and then look in this 8possibly
> surounding pixels which one is present. when this is done, keep  
> iterating
> until the loop is closed (ie, all pixels around are blank or already  
> in the
> ROI).
>
> The only problem is that perhaps in the future some objects may have  
> holes
> in them for instance.
>
> My need for ROI, is mainly for display purposes, although perhaps,  
> in the
> future, it might be used to skip some steps for the image analysis.
>
> Any help would be most appreciated!
>
> Best,
>
> Ghislain
> --
> View this message in context: http://n2.nabble.com/Convert-a-set-of-points-into-an-ROI-tp2864581p2865999.html
> Sent from the ImageJ mailing list archive at Nabble.com.

Karsten
[hidden email]