This project is an implementation for Background Subtract based on GMM model, coded in Python language.
Here we use Test Images for Wallflower Paper to train our background model and test the subtract result.
To improve the subtract result, we use Mathematical Morphology to remove noise and reconnect the disconnected component.
- Runnable Python source files are singleChannel.py and multiChannels.py, each of which is implemented on Gray Scale and RGB Scale. JUST CLONE THE REPOSITORY AND RUN IT!
- Dataset is in WavingTrees directory, which contains two subdirectories which are background_train for model training and person_in for model testing.
- Models which have already been learned with different parameters are in models_learned directory, so that you can move quickly to testing stage.
- Background Model Learning
- Foreground Person Detecting
- [1] Stauffer C , Grimson W E L . Adaptive background mixture models for real-time tracking[C]// IEEE Computer Society Conference on Computer Vision & Pattern Recognition. IEEE Xplore, 2007.
LeoHao (XMU-CS)
2020.10.22