Build under Linux

[lucasa]

Hi!

Do you has any tip how to build opensynther and dependencies under Linux (gcc, etc) ?

Thanks in advance
Lucas Alberto

[dddExperiments]

Hi Lucas,

I'm sorry: there is no easy way of compiling OpenSynther under linux.
BTW OpenSynther is not finished but the only interesting feature is the linear unstructured matching solution.
Why are you interested in compiling it?

Henri

[lucasa]

Hi Henri,
I want to develop a webservice to do automatic detection of robust features
on frames of a video and calculate the 3d positons of the camera. What are
the parts of the code that I can use?

Lucas

2012/9/18 Henri Astre notifications@github.com

Hi Lucas,

I'm sorry: there is no easy way of compiling OpenSynther under linux.
BTW OpenSynther is not finished but the only interesting feature is the
linear unstructured matching solution.
Why are you interested in compiling it?

Henri

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/1#issuecomment-8642822.

[dddExperiments]

Well I guess you're thinking about my pose estimator web service demo. But I don't have open source the web service neither the pose estimation solution. OpenSynther can be used to produce a sparse point cloud from a previous sfm result (photosynth for example). But the triangulation implementation in OpenSynther is not robust (some filters are missing). BTW IMO if you want to track features in a video you should try a KLT based solution...

If you are interested in 6 DOF tracking you may be interested by this paper too.

[lucasa]

Thanks for the answer. i'm going to use the libmv (used in blender for
camera tracking) that has some very good high level functions. My plan is
to use genetic algorithms or other optimization technic to help the
tracking work and get better results.

Lucas

2012/9/20 Henri Astre notifications@github.com

Well I guess you're thinking about my pose estimator web service demohttp://www.youtube.com/watch?v=vFVwmLucyb4.
But I don't have open source the web service neither the pose estimation
solution. OpenSynther can be used to produce a sparse point cloud from a
previous sfm result (photosynth for example). But the triangulation
implementation in OpenSynther is not robust (some filters are missing). BTW
IMO if you want to track features in a video you should try a KLT based
solution...

If you are interested in 6 DOF tracking you may be interested by this
paperhttp://research.microsoft.com/en-us/um/redmond/groups/ivm/ImageBasedLocalization/too.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/1#issuecomment-8716955.