C++ implementation of a background subtraction algorithm, following the method "Adaptive background mixture models for real-time tracking" by Stauffer and Grimson.
The paper can be found here.
This has been tested on WSL (with Ubuntu). It uses OpenMP to process the pixels in parallel.
To use:
- create a
build
folder cmake ../
cmake --build .
./backgroundSubtraction videoname K lr T
, specifying values for videoname, K, lr and T.
For the test video (that can be found here), with the following parameters: