[vtkusers] Data driven VTK

[Juan Santos]

Hi,

I'm fairly new to VTK and I have a question regarding the VTK architecture.

Basically I have a specialized image reconstruction library "like" VTK but
it is data-driven. Basically when data goes in, the processed data comes out
of the pipeline. I understand that VTK works the other way around. When data
is required at the output, it backpropagates that request until it finds
which portions of the pipeline need to be updated.

I would like to connect both pipelines (data-driven at the input of VTK) but
I'm having some trouble. It would be ideal for me if the whole pipeline is
data-driven, but it would be also fine if the VTK portion remains
demand-driven, as long as it gets updated when new data arrives from the
"data-driven" portion. I looked at video capture examples but unfortunately
they relay on the VTK event loop. The data driven portion of the pipeline
already has an event loop (to react to new data at the input) so this is not
compatible with just mark the data as "Modified" and wait for the pipeline
to be updated (as I understand how VTK works). What does work though is if I
manually call the render method of the visualizer after I feed the VTK
pipeline. Unfortunately, that solution its a little bit odd as the "data
conversion" object (the one who connected the data-driven pipeline with VTK)
will need to know about the last object in the pipeline to call the update
method (maybe this is the only way).


Is there a good solution to this kind of problem? I looked into the
"pipeline executive" architecture but I don't know enough of that and it
seems that the only propagation methods available are demand-driven.

I'll appreciate if someone who has some experience with this could point me
in the right direction.

Thanks

Juan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20071210/ee7cc880/attachment.htm>