[Insight-developers] itkImageSeriesReader Bug (contd.)

[Hauke Heibel]

Hi,

 

I added a fixed version to the according bug report posted by my colleague.
The implementation passed all current IO regression tests.

 

It can be found here: http://www.itk.org/Bug/view.php?id=4559

 

- Hauke

 

From: insight-developers-bounces+heibel=cs.tum.edu at itk.org
[mailto:insight-developers-bounces+heibel=cs.tum.edu at itk.org] On Behalf Of
Hauke Heibel
Sent: Tuesday, October 23, 2007 11:37 AM
To: insight-developers at itk.org
Subject: [Insight-developers] itkImageSeriesReader Bug (contd.)

 

Hi again,

 

After some discussions with my colleague we think that we need a little bit
of help and discussion within the community.

 

We understand that currently loading a series of 3D volumes of full
dimensionality via the itkImageSeriesReader into a 3D volume is not
supported.  The system will correctly throw an exception from
itkImageSeriesReader.cxx line 299 (CVS 10/23/2007).

 

In our opinion there is a big issue since before the exception is thrown two
out of bounds memory write operations are performed. The first one occurs on
itkImageSeriesReader.cxx line 158 and the second one in line 176 (again CVS
10/23/2007).

 

The error is not recognized in the first instance, since the system will
only complain sometimes.

 

I tried to come up with a way to load 3D volumes into a single 3D volume in
my prior post on this mailing list
(http://www.itk.org/mailman/private/insight-developers/2007-October/009695.h
tml) but I oversaw the memory access error that already existed. The example
I attached to this e-mail will allow you to debug this issue. I created two
DICOM images myself, wrote them to the HD and read them back via the series
reader.

 

I personally would like very much to see the functionality of being able to
stack together multiple 3D volumes into a single volume via the series
reader and I will investigate the problem further on my side to come up with
a solution that will work in all cases. In case this is not meant to be
supported in the future it is probably still important to fix the memory
access error.

 

We are looking forward to your opinions.

 

Best regards,

Hauke

---

Tim Hauke Heibel

Computer Aided Medical Procedures

Tel: +49.89.289.19403

Fax: +49.89.289.17059

Room: MI 03.13.058

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.itk.org/mailman/private/insight-developers/attachments/20071023/18dbb844/attachment.htm