Hi, In addition to what Matt already said, we harmonized Global Reinit and Reinit. The Reinit behavior was rather unpredictable before as a "random" node from multiple selected nodes was chosen to do a Reinit on it. It wasn't possible to Reinit a set of selected nodes. However, there is a little bit more regarding the logic of the classic Reinit as it also handled the visual rectification of askew images and the like. To keep this functionality as it always was, there is the special case of a single selected image.
The commit with the changes is https://phabricator.mitk.org/rMITK6a8e85861ddb36cbe7d1c619f762721c430533cf Best, Stefan -----Original Message----- From: Daniel Maleike [mailto:c...@maleike.de] Sent: Donnerstag, 14. Juni 2018 09:00 To: mitk-users@lists.sourceforge.net Subject: Re: [mitk-users] "Reinit" behavior in Data Manager: why are images handles in a special way? Hi Matt, that's a good point - thinking about file formats. I didn't think of that because my workflow was a little different: load/generate e.g. vtkPolyData, then manipulate its MITK geometry to translate/rotate the object (those manipulations are respected by rendering). Then I tried reinit and discovered that the transformation defined in MITK's geometry is just ignored there. While I am still not sure if the current behavior is what everybody would expect, your point gives a good explanation for many applications. Thanks! Regards, Daniel > "Clarkson, Matt" <m.clark...@ucl.ac.uk> hat am 13. Juni 2018 um 21:06 > geschrieben: > > > Hi there, > > Im guessing that images have the orientation of their axes specified in the > header, which can be different from the global x,y,z axes. Surfaces and > points dont. So, for example, an mitk::Surface contains a vtkPolyData, and > all coordinates are in x,y,z coordinates. So, there is no other coordinate > axes to chose from. Historically, mitk::PointSet used to just contain x,y,z > coordinates but a while back, around 2015 I think, the file format itself was > updated to include the orientation. > > M > > > > On 13 Jun 2018, at 14:45, Daniel Maleike <c...@maleike.de> wrote: > > > > Hi, > > > > I just noticed that the "Reinit" of Data Manager implements alignment to > > object axes only in the special case of a single selected _image_. I > > understand that for multiple selections a world-axes aligned bounding box > > is more what we expect. > > However, I have a doubt why we handle images different from meshes or point > > clouds. If those objects have their proper coordinate system, why shouldn't > > we align to this system if the user selects explicitly one of those objects? > > > > Can anybody explain / remember the reason for this choice? > > > > Thanks, best regards, > > Daniel > > > > ------------------------------------------------------------------------------ > > Check out the vibrant tech community on one of the world's most > > engaging tech sites, Slashdot.org! http://sdm.link/slashdot > > _______________________________________________ > > mitk-users mailing list > > mitk-users@lists.sourceforge.net > > https://lists.sourceforge.net/lists/listinfo/mitk-users > ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ mitk-users mailing list mitk-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mitk-users ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ mitk-users mailing list mitk-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mitk-users
