New to Libmv

[Aniket]

Hey there, I am looking forward to contribute to this fascinating project. Can someone please get me started and allocate some work. I am fluent in C/C++ and currently enrolled in final year of college. Thanks in advance.

Aniket

[Julien Michot]

Hey Aniket,

You're welcome to contribute to libmv ;)
The first thing to do is to get the source code from our git repository (https://github.com/libmv/libmv), compile it and enjoy.
Do you have some background in computer vision/structure-from-motion?
For the task allocation, Keir has probably some ideas. //TODO(keir) insert some ideas here.

Enjoy,
Julien

On Tue, Dec 13, 2011 at 3:37 PM, Aniket <atn...@gmail.com> wrote:

Hey there, I am looking forward to contribute to this fascinating project. Can someone please get me started and allocate some work. I am fluent in C/C++ and currently enrolled in final year of college. Thanks in advance.

Aniket

--
You received this message because you are subscribed to the Google Groups "libmv-devel" group.
To view this discussion on the web visit https://groups.google.com/d/msg/libmv-devel/-/ubMvSpYoLxEJ.
To post to this group, send email to libmv...@googlegroups.com.
To unsubscribe from this group, send email to libmv-devel...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/libmv-devel?hl=en.

[Keir Mierle]

Hi Aniket,

Recently I've slacked off on my libmv leadership duties in the rush to get motion tracking in Blender working great. That's mostly settled now that Blender 2.61 is in code freeze; I only have 1 big Blender patch left (targeting Blender 2.62).

My libmv task for this week is to reboot the public release of libmv with all the changes I've made in libmv for Blender (better tracking, reconstruction, etc), in a new stripped down package that has much less of what we have in libmv now.

With that in mind, what are your particular interests? Probably the most important bit is how much vision experience you have. If you have vision experience, I can give you projects that require more mathematical sophistication. If not, there are infinite cleanup possibilities.

Keir

[Pierre Moulon]

Hi Aniket,

Perhaps for the computer vision side,

We could take a look to N View triangulation at least square for first contribution.

PS for libmv users:

SSBA was updated... to 3.0.

"News for SSBA-3.0: 
 Added a sparse LM implementation (struct NLSQ_LM_Optimizer and friends) for non-bipartite problems. This can be e.g. useful when 3D lines are added to the sparse reconstruction, or when (soft) constraints between unknowns or priors for latent variables need to be incorporated. Several least-squares terms in the cost functions are also allowed."

Regards,

Pierre

2011/12/13 Keir Mierle <mie...@gmail.com>

[Aniket Handa]

Thanks all for the warm welcome and a quick response.

I have worked on topics like Optical flow using the Lucas-Kanade
(OpenCV), in which we estimated depth from parallax clues, but I would
call myself a beginner.

I would certainly like to take the “more mathematical sophistication”
way as at the end of the day I want learn something out of all this
:-)

N-view triangulation at least squares look like an interesting thing
to work upon, thought I am sure I need to read about it a lot before
some actual contribution.

Regards,

Aniket