[Insight-developers] RE: Yesterday's changes

[Lorensen, William E (Research)]

It seems that you have fixed a problem with the filter. No need to ask the
list about that. Sometimes our baselines are wrong.

-----Original Message-----
From: Mark Foskey [mailto:mark_foskey@unc.edu]
Sent: Tuesday, April 22, 2003 5:41 PM
To: Lorensen, William E (Research)
Cc: insight-developers@public.kitware.com
Subject: Re: Yesterday's changes


Note to the list: I've been changing code related to the Danielsson 
filter, and it's been causing problems.  I'm pulling the list in to the 
discussion with Bill over this.

Lorensen, William E (Research) wrote:
 > There seems to be a 1 pixel outline around the resulting image.
 >
...OK, that was my fault.  (Well, obviously...)

I modified the behavior of the filter and didn't tell anybody.  I 
thought that filter didn't yet have a comparison against a baseline, 
and I modified it to produce the distance filter for the whole image, 
rather than for all of the image except for a one-pixel boundary, as it 
did previously.  So it was the baseline image that had the one-pixel 
boundary.  This boundary was causing problems for some of our research 
projects, and I saw a way to modify the ReflectiveImageRegionIterator 
to get the distance for the whole image.

Actually, the [0, 0] pixel in the image is still incorrect, I have 
realized later this afternoon, although all the others look to be 
correct.  I intended to fix that tomorrow.

I have Cc'ed the list on this, since perhaps I should have written the 
list about this to begin with.  If people object to the whole proposal 
we can revert my changes.


> -----Original Message-----
> From: Mark Foskey [mailto:mark_foskey@unc.edu]
> Sent: Tuesday, April 22, 2003 2:47 PM
> To: Lorensen, William E (Research)
> Subject: Re: Yesterday's changes
> 
> 
> I have run the tests on ReflectiveImageRegionIteratorTest, 
> DanielssonDistanceMapImageFilterTest, and 
> HausdorffDistanceImageFilterTest.  The tests passed, and I have checked 
> in the changes.
> 
> Lorensen, William E (Research) wrote:
> 
>>Mark,
>>Did you run the ReflectiveImageRegionIteratorTest after your changes
>>yesterda? It is not passing on any platforms. Also, related tests are
>>failing.
>>
>>Bill
> 
> 
> 


-- 
Mark Foskey    (919) 843-5436  Computer-Aided Diagnosis and Display Lab
mark_foskey@unc.edu            Department of Radiology, CB 7515, UNC
http://www.cs.unc.edu/~foskey  Chapel Hill, NC  27599-7515