*Update: 1.5
-Heterogeneous Tracking Ability (Make and track multiple models)
-Parallel Processing optimization (Speed up your tracking)
-Color-ID Potential (Can try to determine color-coded ID)
-Trackfixer (A simple utility for making quick improvements to data)
*Important: to use the new version of Multitrack, now called Manytrack, you need to put your model files into their own separate folder.
Download the Biotrack Pack
This is the first iteration of a modular software suite for tracking and analyzing multiple agents. Licensed under GPL it is free for non-commercial use and development. This open-source project makes use of the OpenCV, PointCloud Library, and QT libraries.
Download Biotrack Pack for:
- Ubuntu, Ubuntu(64 Bit) (supported)
- Mac (64 bit) (experimental)
- Source Code is available at Github: https://github.com/biotracking/biotrack (source)
We have updated sample files to practice using the slightly new tracking system: Biotracking Samples (Sample files provided by Georgia Tech, Kirsten Traynor, Zachary Shaffer, and Stephen Pratt at ASU)
Biotrack Pack development is led by PhD student Andy Quitmeyer and Prof. Tucker Balch with significant help from Brian Hrolenok. It was created with support from NSF Award 0960618 and ONR contract 550740. Matthew Flagg and Prof. James M. Rehg contributed to an earlier version of this software.
Our tracking approach is based on:
The Multi-Iterative Closest Point Tracker: An online algorithm for tracking multiple interacting targets. Journal of Field Robotics. Adam Feldman, Maria Hybinette, Tucker Balch. (2012).
*We have now modified this approach to include color data in the tracking algorithm.
Testimonial
Stephen C. Pratt, Assistant Professor Arizona State University
I’ve had a chance to play with the new suite a bit, and it’s very nice! The interface is clean and easy to follow. The visual cues are very helpful, and its good to be able to toggle them. The Backgrounder and ModelMaker tools worked very well, and are a big help. Saving settings is fantastic, and makes tuning much more efficient. I went through all of the videos and wiki instructions and then tested everything by running a fresh video. After a bit of futzing with the parameters, it worked well. There is still some track stealing, but I did not necessarily optimize the parameters, so perhaps that could be reduced. Everything went pretty quickly, except for actual tracking, but that’s OK since it can run unsupervised.
Getting more information
Our Wiki provides full examples and tutorials for obtaining and using the included software:
https://github.com/biotracking/biotrack/wiki
If you would like to receive updates about our software, please subscribe your email address in the box below:
| Subscribe to the bio-tracking.org announcement list |
| Email: |
| Visit this group |
To examine others’ questions about using the software, find fixes for bugs, or discuss how the software works, join the web-forum/email list:
https://groups.google.com/a/bio-tracking.org/forum/?fromgroups#!forum/discuss
alternately you can email your questions, or topics directly to:
discuss|at|bio-tracking.org
If you have other questions not covered in the above wiki, you can direct any questions to:
info|at|bio-tracking.org
To report a bug in the program please send a detailed description of your bug to:
bug-report|at|bio-tracking.org
   |