[vtk-developers] New image mappers

David Gobbi david.gobbi at gmail.com
Thu Mar 31 15:22:55 EDT 2011


A few new image mapper items have just been committed.  I'd love to
hear some feedback about these classes (bug reports, feature requests,
praise, whatever).  So far I haven't heard from anyone who has tried
them since I submitted them.

The new items:

1) A "vtkImageStack" class and a SetLayerNumber() method in
vtkImageProperty.  The image stack layers the images in the correct
order, and it does a simple two-pass rendering trick where it delays
writing to the depth buffer until all of the images have rendered, in
order to avoid depth-buffer coincidence artifacts.

2) Assemblies can now be used with vtkImageSlice.  This was broken before.

3) Checkerboarding is supported, it can be turned on and off in the
image property.

These are probably the last changes that I will be making for the next
few weeks.

 - David


On Mon, Mar 21, 2011 at 11:27 AM, David Gobbi <david.gobbi at gmail.com> wrote:
> The dashboard is coming up clean for the new image mapper classes
> (vtkImageSliceMapper and vtkImageResliceMapper) so I'm confident
> that they're ready to use now.
>
> There are two issues remaining to be fixed:
> - They always render duing the "Opaque" pass
> - They don't work properly when in an assembly
>
> I'm thinking that I'll let the user select which pass to render in:
> 1) as an underlay (render first, don't set the depth buffer)
> 2) as per the default  (check and set the depth buffer)
> 3) in "translucent" pass (check but don't set the depth buffer)
> 4) as an overlay (render last, don't check or set depth buffer)
>
> Not giving the user access to all of these options seems unnecessarily
> restrictive (and this restrictiveness was one thing I disliked about the
> vtkImageActor).
>
> The assembly issue is a bug, so I just have to fix it.  I'm also strongly
> considering an image-specific assembly subclass called "vtkImageStack"
> to simplify layering of images.  The idea is that each image can be
> assigned a layer number, and vtkImageStack will ensure that the images
> in the stack are sorted according to this number.  This will also allow me
> to add some tricks to eliminate all depth buffer coincidence problems
> between the images.
>
>  - David
>
>
> On Tue, Mar 15, 2011 at 8:25 AM, David Gobbi <david.gobbi at gmail.com> wrote:
>> The new image mappers have been merged into VTK.
>> For a description of these mappers, see the wiki page:
>> http://www.vtk.org/Wiki/VTK/Image_Rendering_Classes
>>
>>  - David
>>
>>
>> On Mon, Mar 14, 2011 at 7:40 AM, David Gobbi <david.gobbi at gmail.com> wrote:
>>> I just pushed another image-mapper patchset to gerrit this
>>> morning.  In this patchset, vtkImage has been renamed to
>>> vtkImageSlice and more bugs have been fixed.  Unless
>>> something comes up, I'll merge this into VTK tomorrow.
>>>
>>>  - David
>>>
>>>
>>> On Fri, Mar 11, 2011 at 10:10 AM, David Gobbi <david.gobbi at gmail.com> wrote:
>>>> I've updated gerrit with my latest image mapper patches.
>>>> http://review.source.kitware.com/1163
>>>> This patch set includes a full set of regression tests.
>>>> I'm hoping to be able to move this into VTK next week.
>>>>
>>>>  - David
>>>>
>>>>
>>>> On Tue, Mar 8, 2011 at 2:06 PM, David Gobbi <david.gobbi at gmail.com> wrote:
>>>>> I've pushed my new image mapper classes to Gerrit,
>>>>> you should be able to find them at the following link:
>>>>> http://review.source.kitware.com/1163
>>>>>
>>>>> They come with one example (in python, of course)
>>>>> and over the next few days I'll add some regression
>>>>> tests.
>>>>>
>>>>>  - David
>>>>>
>>>>
>>>
>>
>



More information about the vtk-developers mailing list