[Paraview] Newbie: contributing with an implementation of the Splatting algorithm

[Jordi Campos i Miralles]

Hi all,

I'm a PhD student [0] at Universitat Politècnica de Catalunya [1] and
Universitat de Barcelona [2], working on Volume Visualization at MoViBio
research group [3].

The research group has its own software system Hipovis [4], but I am
evaluating the possibility of using other software systems. Right now,
I'm evaluating using ParaView/VTK to test current research approaches on
Focus+Context Volume Visualization based on the Splatting algorithm [5].

As far as I know, VTK libraries does not offer the Volume Splatting
algorithm [6] and there are no available examples of implementing it
using VTK. However, I found references to works that use VTK to
implement Splatting approaches [7], but no code is available.



Hence, I have the following questions::::::::::::::::

A- Is there any VTK's architectural reason why the splatting algorithm
does not suit the VTK libraries? May be implementing the splatting
algorithm inside the VTK's pipeline architecture will not provide enough
performance to develop research activities, will it?

B- Although ParaView offers fast volume rendering (textures), precise
v.r. (RayCast) and both qualities v.r. (using VolumePro hw), will it be
worth if I work on adding the Splatting algorithm and Focus+Context
features? (for me it is worth because it is the basis of my thesis but I
would like to know if there are other people also interested in it).

C- If I want to be involved in VTK development to implement the
Splatting algorithm, which steps do you recommend me? From you
experience, how long does it take to a PhD student (with C++ and
graphics knowledge) to get involved with the project and develop new
code? (I know this is a generic question, but I need some estimation). I
already bought the two books [8][9], reviewed the coding standards [10]
and subscribed to the list.

D- If I want to be involved in ParaView development to use the Splatting
algorithm and implement Focus+Context features, which steps do you
recommend me?  From you experience, how long does it take to a PhD
student (with C++ and graphics knowledge) to get involved with the
project and develop new code? (I know this is a generic question, but I
need some estimation). I already ordered the book [11] and subscribed to
the list.

E- Which are the good and bad points of ParaView/VTK versus VolumeShop
[12]? It seems VolumeShop is not really open source (and we are looking
an open source project to get involved), it is mainly used on MsWindows
platforms (and we are interested in a multi-platform environment), and
it does not provide parallel rendering.

F- Which are the good and bad points of ParaView/VTK versus VolView
[13]?




I would like to known your opinion about these questions and any other
possible important factors you think I should pay attention.



Thank you very much in advance.


jor;)i



PD: I already have sent a similar post to VTK's list, specifically about
VTK's issues.


References:
[0] http://truja.lsi.upc.edu/movibio/Members/jordi
[1] http://www.upc.es
[2] http://www.ub.edu
[3] http://truja.lsi.upc.edu/movibio
[4] http://truja.lsi.upc.edu/movibio/ResearchLines/soft
[5] Lee Westover: Interactive volume rendering, 1989
[6]
http://public.kitware.com/pipermail/vtkusers/2001-January/055005.html
[7] Zwicker, Pfister, Baar and Markus Gross: EWA volume splatting, 2001
[8] The Visualization Toolkit An Object-Oriented Approach To 3D Graphics
[9] VTK User's Guide
[10] http://www.vtk.org/Wiki/VTK_Coding_Standards
[11] ParaView User's Guide
[12] http://www.cg.tuwien.ac.at/volumeshop
[13] http://www.volview.com