Download the data:
In order to track the use of this dataset, we ask that you first email us, and we will respond with the download link (and, try to coordinate passing you the most up-to-date related code). To get the link, email David Meger at dpmeger(at)cs(dot)ubc(dot)ca.
For your information, several sample images are available:
and here is an example of one associated XML file that describes the image geometry and annotations:
ALPHA NOTE
Please note that very few users outside UBC have so far started using this code and data. We are actively trying to make it as easy and stable for others as possible, but there is a good chance you will find some non-robust aspects of the code, in particular. Please send feedback to dpmeger@cs.ubc.ca and someone at UBC will try to address your issue as soon as possible. Also note that we welcome collaboration, and if you begin to use the data extensively, we are happy to add you as a team member to contribute more robust tools to work with this data.
Instructions to start Using:
- Untar the data somewhere convenient.
- Download the python helper functions witht the command: svn co https://ubc-ros-pkg.svn.sourceforge.net/svnroot/ubc-ros-pkg/branches/ubc_vrs_public_version
- Obtain dependency python packages: python-opencv, python-wxgtk2.8, python-configobj
- Edit the config file at: ubc_vrs_public_version/config/vrs_config.ini:
- Point the root_path to where you downloaded the data.
- Point the output_dir to where you want to store images and temporary files you may want to keep (writtable).
- Point the graphics_cache to a writeable directory where temp files you dont want to keep will be stored. (can leave /tmp unless it's deleted regularily.)
- Run one of the sample programs, such as the annotators within the ubc_vrs_annotation_tools sub-directory or the results viewer within ubc_vrs_results_analysis.
- Enjoy!