Re: [ros-users] Object recognition + positioning package?

Hi Raphael,
> Do you recommend any minimal size for the image DB?
> I am having trouble (actually runtime errors) while using the clustering
> mode.
The number of images and the number of extracted SIFT features on
average per image are the ones that basically directly influence the
parameters tree_k and tree_levels. I unfortunately do not have
anything else but a highly intuitive guess that 20 images with 500
SIFT features on average shall go in the database with (tree_k,
tree_levels) = (3, 3) or (4, 4). I used (6, 6) for ~150 images with
500 SIFT features.


> I will dig in the problem today but do you think there any chance this is
> linked to my image DB being too small (around 20 pictures)?
> In other words: is it just me? :)
The size of the DB shall not give runtime errors.

> Finally, do you know a kind of automatized way to import the  germandeli/
> http://www.germandeli.com/  catalogue you advertise on the wiki? So far I've
> been manually saving the thumbnails from the website...
Yes, we do. Please a have look into germandeli_crawler package, you
will find a README.txt which will tell you how to run it.


cheers and let me know if you get into more troubles,
D.
> View this message in context: http://ros-users.122217.n3.nabble.com/Object-recognition-positioning-package-tp2399294p2450387.html
> Sent from the ROS-Users mailing list archive at Nabble.com.
> _______________________________________________
> ros-users mailing list
> ros-users@code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users
>



--
MSc. Dejan Pangercic
PhD Student/Researcher
Intelligent Autonomous Systems Group
Technische Universität München
Telephone: +49 (89) 289-26908
E-Mail: dejan.pangercic@cs.tum.edu
WWW: http://ias.cs.tum.edu/people/pangercic