Multi-object tracking using min-cost flow

This is a simple Python implementation of tracking algorithm based on global data association via network flows [1].

Targets are tracked by minimizing network costs built on initial detection results.

Dependencies

• numpy
• OpenCV (for image reading, processing)
• ortools (for optimizing min-cost flow)

Usage

Please modify test.py and mcftracker.py to adapt your tracking targets. You can test this implementation as:

% python test.py

To include it in your project, you just need to:

tracker = MinCostFlowTracker(some_parameters)
tracker.build_network(images)
optimal_flow, optimal_cost = tracker.run()

You can use fibonacci search to reduce computation costs.

License

MIT

References

[1] L. Zhang et al., "Global data association for multi-object tracking using network flows", CVPR 2008