MantisBT - ParaView
View Issue Details
IDProjectCategoryView StatusDate SubmittedLast Update
0012509ParaView(No Category)public2011-08-15 07:382016-08-12 09:58
ReporterPaul Melis 
Assigned ToKitware Robot 
PrioritynormalSeverityminorReproducibilityhave not tried
StatusclosedResolutionmoved 
PlatformOSOS Version
Product Version3.10.1 
Target VersionFixed in Version 
ProjectTBD
Topic Name
Typeincorrect functionality
Summary0012509: Extract Subset filter strange behaviour
DescriptionI have an image dataset A (1137x716x834) that was created using the
Extract Subset filter from a larger dataset. It is stored in a .pvti +
8*.vti files. The extents of A are

  X: 514-1650 (dim 1137), Y: 455-1170 (dim 716), Z: 0:833 (dim 834)

I want to create a subsampled version of A by once again using the
Extract Subset filter, with as VOI the full extents above and setting
its sample rates to 2. But to my surprise the subsampled dataset (using
PV 3.10.1 in standalone mode, btw) ends up having extents

  X: 514-825 (312), Y: 455-584 (130), Z: 0-416 (417)

and covers only the upper portion of A, not the whole set. I also don't
understand where the weird reduction in X and Y dimensions comes from
(factors 3.6 and 5.5, resp). See the attached screenshots.
Steps To Reproduce
Additional Information
TagsNo tags attached.
Relationships
Attached Filesjpg subset0.jpg (120,540) 2011-08-15 07:38
https://www.vtk.org/Bug/file/9001/subset0.jpg
jpg

jpg subset1.jpg (95,295) 2011-08-15 07:38
https://www.vtk.org/Bug/file/9002/subset1.jpg
jpg

jpg subset2.jpg (117,303) 2011-08-15 07:38
https://www.vtk.org/Bug/file/9003/subset2.jpg
jpg
Issue History
Date ModifiedUsernameFieldChange
2011-08-15 07:38Paul MelisNew Issue
2011-08-15 07:38Paul MelisFile Added: subset0.jpg
2011-08-15 07:38Paul MelisFile Added: subset1.jpg
2011-08-15 07:38Paul MelisFile Added: subset2.jpg
2012-01-26 23:51yifei wuNote Added: 0027984
2012-01-26 23:59yifei wuNote Edited: 0027984bug_revision_view_page.php?bugnote_id=27984#r416
2012-01-27 07:39Paul MelisNote Added: 0027985
2012-01-27 07:55yifei wuNote Added: 0027986
2012-01-27 07:56yifei wuNote Edited: 0027986bug_revision_view_page.php?bugnote_id=27986#r418
2012-02-03 06:04Paul MelisNote Added: 0028023
2016-08-12 09:58Kitware RobotNote Added: 0038041
2016-08-12 09:58Kitware RobotStatusbacklog => closed
2016-08-12 09:58Kitware RobotResolutionopen => moved
2016-08-12 09:58Kitware RobotAssigned To => Kitware Robot

Notes
(0027984)
yifei wu   
2012-01-26 23:51   
(edited on: 2012-01-26 23:59)
The extraction is actually working as intended. It does not interpolate. From documentation at
http://www.paraview.org/ParaView3/Doc/Nightly/html/classvtkPVExtractVOI.html [^]

Set the sampling rate in the i, j, and k directions. If the rate is > 1, then the resulting VOI will be subsampled representation of the input. For example, if the SampleRate=(2,2,2), every other point will be selected, resulting in a volume 1/8th the original size.
(0027985)
Paul Melis   
2012-01-27 07:39   
It would be great if you could show me how the input extents get transformed into output extents based on the VOI and sampling settings. Because the way the sampling rate works shown in my original report doesn't make any sense to me. Either I completely misunderstand how this filter is supposed to work or there is a bug at work here.

Note that I used the full extents as VOI, so in my opinion only subsampling should happen. I set the sampling rate to 2, but the number of values in X, Y and Z is reduced by factors of 3.6, 5.5 and 2. So only the Z sampling rate has been correctly applied. I would expect output dimensions of around 569 for X (312 instead) and 358 for Y (130 instead).
(0027986)
yifei wu   
2012-01-27 07:55   
(edited on: 2012-01-27 07:56)
The second issue is the extents are referencing based on the coordinates of the VOI rather than the global coordinate system.

If you wanted to use the full extents of the VOI, you'd select 0-1137 because 514-1650 goes out of bounds i.e. start at VOI 514 position, not 514 from the overall grid.

1137 - 514 = 623 623/2 ~ 312.

I don't think it's a bug, but the documentation could be augmented for clarity.
(0028023)
Paul Melis   
2012-02-03 06:04   
The extent entry boxes don't allow you to enter any extent outside the input extent. So I can't enter 0-1137 for example, even if I wanted to.

And I'm surprised that global extents are propagated throughout the pipeline, but for a filter you would have to use *local* extents?
(0038041)
Kitware Robot   
2016-08-12 09:58   
Resolving issue as `moved`.

This issue tracker is no longer used. Further discussion of this issue may take place in the current ParaView Issues page linked in the banner at the top of this page.