Weasis - Sorting Instances

362 views
Skip to first unread message

Doug Steinfeld

unread,
Feb 6, 2014, 1:29:36 PM2/6/14
to
My apologies if this has been covered elsewhere.  If so, a pointer would be appreciated!

Weasis 1.2.8 as launched from http://dicom.vital-it.ch:8089/dcm4chee-web3, as well as 1.2.4 launched from our own DCM4CHEE server fails to sort the instances when using 2D-DICOM Viewer -> Sort Stack By commands.  In particular, the affected studies are angio.  Instances appear to be displayed in a random order, possibly the order in which they were downloaded.  Regardless of selection, Sort Stack By appears to do nothing in both cases.  I do not see a Jira ticket open/reopened/in progress on this.  I've tried highlighting all instances in the study before making the selection, but no effect.  Is there some obvious technique I'm missing?

--> more

Whereas with an MR, Weasis will display a thumbnail of each series in the DICOM Explorer window (and will correctly order the instances per the Sort Stack By command), with an angio (and perhaps US as well, see an old thread at http://forums.dcm4che.org/jiveforums/message.jspa?messageID=20519) the series is broken out so that each instance (which is a multiframe) is thumbnailed and displayed separately in the DICOM Explorer window.  It is this set of thumbnails that is misordered (with respect to Instance Number or Acquisition Date/Time) and it doesn't appear that this set can be re-ordered using Sort Stack By.  This makes it very difficult for a reader to view the angio loops in the sequence in which they were scanned.

So again, am I missing something obvious?

Many thanks!
Doug

nicolas...@gmail.com

unread,
Feb 6, 2014, 4:16:26 PM2/6/14
to dcm...@googlegroups.com
Could you provide a sample or give a reference from the demo server?

The default sorting is "Instance Number" => should be the default scanned order. Did you try to sort by "Anatomical Direction". The latter should be the default mode according to the DICOM recommendation but in practice there are many problematic cases.

Doug Steinfeld

unread,
Feb 6, 2014, 5:12:56 PM2/6/14
to dcm...@googlegroups.com
Thanks for your quick reply.  Sure--here's a demo server example:

Patient Name GRUSELAMBIX, XA with 14 instances in 1 series, Study date/time of 10/15/2003 12:00.  Expanding the study to the instance level, 14 instances are ordered by instance number and acquisition date time ascending.  When the "Open Web Viewer" icon is clicked, Weasis downloads the exam showing a single thumbnail  in the DICOM Explorer window.  When the download completes, 14 thumbnails appear in the DICOM Explorer window.  Moving from the top thumbnail down, #1 has an acquisition time of 9:45:43, #2 has an acquisition time of 10:04:24, and #3 has an acquisition time of 9:43:40.  These times correspond to instances 4, 14, and 1.  There does not appear to be any way to order these thumbnails in ascending instance or acquisition time order.

nicolas...@gmail.com

unread,
Feb 7, 2014, 5:57:49 AM2/7/14
to dcm...@googlegroups.com
This is a special case because all the files shares the same Series Instance UID. That's why all the files are first merged in one series and then split as multiframe with different SOP Instance UID, The character '#' defines a sub-series that has been split. Then each series or sub-series are sorted by first the Series Number and if no value or equal by Slice Location, Image Orientation Plane, Series Date an Time. In this case all the DICOM attributes have no value or the same values. 

Adding Acquisition Date and Time should works. I will add this fix in the next release.

biomed.s...@gmail.com

unread,
Sep 6, 2017, 10:52:20 AM9/6/17
to dcm4che
Dear researcher,
Can u email me the dataset@GRUSELAMBIX which u have downloaded from osirix?I need it ffoe my thesis work.I would highly appreciate if u could help me out.
Reply all
Reply to author
Forward
0 new messages