how to access the ROI programmatically?

[Per Christian Henden]

Hi group,

I need to do a check for whether or not a certain point/pixel coordinate is
contained in the ROI.
My code is of the form
for all pixels, if pixels is in ROI, do stuff
The ROI is a path resulting from freehand selection.

I was not able to find such a function in the API documentation. Is there a
function like the one I'm describing? Any suggestions on how to implement it?

Cheers,

PER

[Gabriel Landini]

> I need to do a check for whether or not a certain point/pixel coordinate is
> contained in the ROI.
> My code is of the form
> for all pixels, if pixels is in ROI, do stuff
> The ROI is a path resulting from freehand selection.
>
> I was not able to find such a function in the API documentation. Is there a
> function like the one I'm describing? Any suggestions on how to implement
it?

There is a method in Roi.java called "contains" which perhaps does what you
need (not entirely sure).
However if you have many pixels this may be slow.

Alternatively you could create a binary image that has just the filled ROI,
then use this image to direct the analysis (scan this image and all set
pixels give you the coordinates of the pixels in the original that need to be
processed).

I hope it helps.

Gabriel

[Joachim Wesner]

Yes, there is contains(), and it´s in the docs

http://rsb.info.nih.gov/ij/developer/api/ij/gui/Roi.html (look for
contains)

but it´s very general yet a bit slow, in case you need to access it in a
tight loop you better create that mask Gabriel mentions once like

            byte rMask[] = null;
            Rectangle rRect = null;
            if (roi != null) {
                  rRect = roi.getBounds();
                  ImageProcessor ipm = roi.getMask();
                  if (ipm != null) {
                        rMask = (byte[])ipm.getPixels();
                  }
            }

and later use your own contains() function.

      boolean contains(int ix, int iy, byte[] mask, Rectangle rect) {

            if (rect == null)
                  return true;

            if (!rect.contains(ix, iy))
                  return false;

            if (mask == null)
                  return true;

            return (mask[(iy-rect.y)*rect.width+(ix-rect.x)] != 0);
      }

in a tight loop like

if (contains(ix, iy, rMask, rRect))
      {
// do smothing if (ix,iy) is contained in the ROI
      }

which is much faster!


JW



                                                                                                                                       
                      Gabriel Landini                                                                                                  
                      <G.Landini@BHAM.         An:      [hidden email]                                                            
                      AC.UK>                   Kopie:                                                                                  
                      Gesendet von:            Thema:   Re: how to access the ROI programmatically?                                    
                      ImageJ Interest                                                                                                  
                      Group                                                                                                            
                      <[hidden email]                                                                                                
                      .GOV>                                                                                                            
                                                                                                                                       
                                                                                                                                       
                      03.07.2006 22:20                                                                                                
                      Bitte antworten                                                                                                  
                      an ImageJ                                                                                                        
                      Interest Group                                                                                                  
                                                                                                                                       




> I need to do a check for whether or not a certain point/pixel coordinate
is
> contained in the ROI.
> My code is of the form
> for all pixels, if pixels is in ROI, do stuff
> The ROI is a path resulting from freehand selection.
>
> I was not able to find such a function in the API documentation. Is there
a
> function like the one I'm describing? Any suggestions on how to implement

it?

There is a method in Roi.java called "contains" which perhaps does what you

need (not entirely sure).
However if you have many pixels this may be slow.

Alternatively you could create a binary image that has just the filled ROI,

then use this image to direct the analysis (scan this image and all set
pixels give you the coordinates of the pixels in the original that need to
be
processed).

I hope it helps.

Gabriel



______________________________________________________________________
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email 
______________________________________________________________________

[Anna Barnes]

Dear All
Could I ask for some clarification on how the import dicom sequence  
works.  Does it work on the principle that the *.dcm files are all in  
the order that they are listed in the directory.  The reason I ask is  
because that I notice in unix when using ls it lists the *dcm files  
that start with a 1 first and then list the others therefore  
IM-1000.dcm comes before IM-900.dcm.  Is this the order they will  
stack ?  How can I check ?

Anna


Anna Barnes, PhD
fMRI Research Laboratory
Department of Radiology
Columbia University College of Physicians and Surgeons
710 W168th St
New York, NY10032
USA
[phone]

[Anna Barnes]

I worked out how to check so now how do I fix it to read the file  
names in numerical order ?

Anna

On Jul 3, 2006, at 5:59 PM, Anna Barnes wrote:

> Dear All
> Could I ask for some clarification on how the import dicom sequence  
> works.  Does it work on the principle that the *.dcm files are all  
> in the order that they are listed in the directory.  The reason I  
> ask is because that I notice in unix when using ls it lists the  
> *dcm files that start with a 1 first and then list the others  
> therefore IM-1000.dcm comes before IM-900.dcm.  Is this the order  
> they will stack ?  How can I check ?
>
> Anna
>
>
> Anna Barnes, PhD
> fMRI Research Laboratory
> Department of Radiology
> Columbia University College of Physicians and Surgeons
> 710 W168th St
> New York, NY10032
> USA
> [phone]

[Wayne Rasband]

> Could I ask for some clarification on how the import dicom sequence  
> works.  Does it work on the principle that the *.dcm files are all  
> in the order that they are listed in the directory.  The reason I  
> ask is because that I notice in unix when using ls it lists the  
> *dcm files that start with a 1 first and then list the others  
> therefore IM-1000.dcm comes before IM-900.dcm.  Is this the order  
> they will stack ?  How can I check ?

The File>Import>Image Sequence command opens DICOM files in image  
number (0020,0013) order when all the images are in the same series,  
otherwise it opens them in alphanumeric file name order.

-wayne

[Anna Barnes]

Hi
Thanks for explaining.  So I tried to fix it by creating 2 substacks  
and then adding them back together in the right order but I'm not  
sure how to do it.  Is there away of adding the two new stack  
together so that the images are reordered ?

Thanks
Anna

On Jul 3, 2006, at 6:19 PM, Rasband Wayne wrote:

>> Could I ask for some clarification on how the import dicom  
>> sequence works.  Does it work on the principle that the *.dcm  
>> files are all in the order that they are listed in the directory.  
>> The reason I ask is because that I notice in unix when using ls it  
>> lists the *dcm files that start with a 1 first and then list the  
>> others therefore IM-1000.dcm comes before IM-900.dcm.  Is this the  
>> order they will stack ?  How can I check ?
>
> The File>Import>Image Sequence command opens DICOM files in image  
> number (0020,0013) order when all the images are in the same  
> series, otherwise it opens them in alphanumeric file name order.
>
> -wayne

[William O'Connell]

Hi Per,
Maybe check that roi isn't null, then something like this:


      Roi roi =imp.getRoi();
      Rectangle rect = roi.getBounds();
      rx = rect.x; ry = rect.y; w = rect.width; h = rect.height;
        for(int y=ry; y<ry+h; y++)
          for(int x=rx; x<rx+w; x++) {
            if(roi.contains(x, y)) {
              ......do what you want to do
            }
           
 
 -------------- Original message ----------------------
From: Per Christian Henden <[hidden email]>

> Hi group,
>
> I need to do a check for whether or not a certain point/pixel coordinate is
> contained in the ROI.
> My code is of the form
> for all pixels, if pixels is in ROI, do stuff
> The ROI is a path resulting from freehand selection.
>
> I was not able to find such a function in the API documentation. Is there a
> function like the one I'm describing? Any suggestions on how to implement it?
>
> Cheers,
>
> PER

[Ahmed Khwaja]

Wayne, Something is wrong with the import sequence. I copied and renamed
first two dicom files (09752176 and 09752197) in a directory so that there
will be an additional two files (09769678 and 09769679) after the last file
(09769677) in the same directory. When I use Import>Image Sequence, the two
additional files (09769678 and 09769679) appear just after the first
(09752176) and third image (09752197) in the sequence. Is there any
particular reason this thing is happening?
Ahmed


-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Anna
Barnes
Sent: Monday, 03 July, 2006 7:39 PM
To: [hidden email]
Subject: Re: import dicom sequence plugin

Hi
Thanks for explaining.  So I tried to fix it by creating 2 substacks and
then adding them back together in the right order but I'm not sure how to do
it.  Is there away of adding the two new stack together so that the images
are reordered ?

Thanks
Anna

On Jul 3, 2006, at 6:19 PM, Rasband Wayne wrote:

>> Could I ask for some clarification on how the import dicom sequence
>> works.  Does it work on the principle that the *.dcm
>> files are all in the order that they are listed in the directory.  
>> The reason I ask is because that I notice in unix when using ls it
>> lists the *dcm files that start with a 1 first and then list the
>> others therefore IM-1000.dcm comes before IM-900.dcm.  Is this the
>> order they will stack ?  How can I check ?
>
> The File>Import>Image Sequence command opens DICOM files in image
> number (0020,0013) order when all the images are in the same series,
> otherwise it opens them in alphanumeric file name order.
>
> -wayne

[Per Christian Henden]

Thanks!

Using contains() to build the mask and then just reading the mask to decide if
I want to include the current pixel or not in the computation sounds like a
good idea. To speed things up further I will ignore pixels outside the
bounding rectangle of the ROI completely, i.e. make the for all pixels loop
consider only the bounding rectangle.

Cheers,
PER

On Monday 03 July 2006 22:56, Joachim Wesner wrote:

> Yes, there is contains(), and it´s in the docs
>
> http://rsb.info.nih.gov/ij/developer/api/ij/gui/Roi.html (look for
> contains)
>
> but it´s very general yet a bit slow, in case you need to access it in a
> tight loop you better create that mask Gabriel mentions once like
>
>             byte rMask[] = null;
>             Rectangle rRect = null;
>             if (roi != null) {
>                   rRect = roi.getBounds();
>                   ImageProcessor ipm = roi.getMask();
>                   if (ipm != null) {
>                         rMask = (byte[])ipm.getPixels();
>                   }
>             }
>
> and later use your own contains() function.
>
>       boolean contains(int ix, int iy, byte[] mask, Rectangle rect) {
>
>             if (rect == null)
>                   return true;
>
>             if (!rect.contains(ix, iy))
>                   return false;
>
>             if (mask == null)
>                   return true;
>
>             return (mask[(iy-rect.y)*rect.width+(ix-rect.x)] != 0);
>       }
>
> in a tight loop like
>
> if (contains(ix, iy, rMask, rRect))
>       {
> // do smothing if (ix,iy) is contained in the ROI
>       }
>
> which is much faster!
>
>
> JW
>
>
>
>
>                       Gabriel Landini
>                       <G.Landini@BHAM.         An:      [hidden email]
>                       AC.UK>                   Kopie:
>                       Gesendet von:            Thema:   Re: how to access
> the ROI programmatically? ImageJ Interest
>                       Group
>                       <[hidden email]
>                       .GOV>
>
>
>                       03.07.2006 22:20
>                       Bitte antworten
>                       an ImageJ
>                       Interest Group
>
> > I need to do a check for whether or not a certain point/pixel coordinate
>
> is
>
> > contained in the ROI.
> > My code is of the form
> > for all pixels, if pixels is in ROI, do stuff
> > The ROI is a path resulting from freehand selection.
> >
> > I was not able to find such a function in the API documentation. Is there
>
> a
>
> > function like the one I'm describing? Any suggestions on how to implement
>
> it?
>
> There is a method in Roi.java called "contains" which perhaps does what you
>
> need (not entirely sure).
> However if you have many pixels this may be slow.
>
> Alternatively you could create a binary image that has just the filled ROI,
>
> then use this image to direct the analysis (scan this image and all set
> pixels give you the coordinates of the pixels in the original that need to
> be
> processed).
>
> I hope it helps.
>
> Gabriel
>
>
>
> ______________________________________________________________________
> This email has been scanned by the MessageLabs Email Security System.
> For more information please visit http://www.messagelabs.com/email
> ______________________________________________________________________