Implementation and comparison of robust PCA methods, with especial focus on their utility in image cleaning (removal of specularities, non-Lambertian effects) and background/foreground separation. Currently implements:
- RPCA (using accelerated proximal gradient)
- RPCA (using exact ALM)
- RPCA (using inexact ALM)
- Fast PCP [Rodriguez,Wohlberg]
The following incremental robust PCA methods are also implemented:
- Online RPCA [Feng,Xu]
- Incremental Fast PCP [Rodriguez,Wohlberg]
- iFrALM
Current dependencies:
- Numpy
- Scikit-learn
- Scikit-image
- Scipy
- Matplotlib
- Cvxpy