[Soc-2013-dev] Weekly Report #6 Motion Tracking

[Joseph Mansfield]

=This week=

This week I worked on figuring out how multiview reconstruction should be
implemented on the libmv side and I've been making some changes to the C
API to accommodate it. So far, this has included updating the Tracks class
to have tracks across different views and unifying some of the libmv
interface. My thinking is that single view reconstruction shouldn't be
treated as a separate case to multiview reconstruction - it's simply when
data for only one view has been given. To move in this direction, I've
unified the existing modal solving functionality with the non-modal
solving. Now there's just a nice clean libmv_solve function.

I also have a non-functional patch from the beginning of the week that I'd
like to submit to trunk, but I don't want to do that until it's been
reviewed. It just tidies up the C API (some function renaming, spelling
corrections, etc.).

I've also been updating my GSoC wiki page with documentation. The plans for
multiview reconstruction have been partly written up. I have a backlog of
weekly updates that I should get on there soon.

=Next week=

I plan to have a working implementation of the libmv side of multiview
reconstruction. I've got a reasonable idea of how this should work.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.blender.org/pipermail/soc-2013-dev/attachments/20130727/63a30b3a/attachment.htm