Overview

This library containing 7 popular background subtraction algorithms:

• adaptive median filtering
• eigenbackground
• single Gaussian
• Gaussian mixture models
• adaptive Gaussian mixture models
• running mean
• mediod filtering

This library makes use of OpenCV.

Building with CMake

$ mkdir _build
$ cd _build
$ cmake ..
$ make

Building Python Interface

1. Install OpenCV with Python bindings enabled.

2. Edit include_dirs in the setup.py.

3. Build the cython module:

    $ pip install numpy cython
    $ python setup.py build_ext --inplace
   

4. Test the module using stream from the webcam:

    $ python pybgs_test.py
   

Citation

If you find this software useful, please consider citing:

• Parks DH, Fels SS. 2008. Evaluation of background substraction algorithms with post-processing. IEEE Fifth International Conference on Advanced Video and Signal Based Surveillance (AVSS2008): 192-199.

License

This software is released under GNE General Public License version 3. See LICENSE.txt for further details.

Contact

Inquires regarding this software can be directed to: donovan_parks_at_gmail_dot_com