Python implementation of SSC-OMP and SSC-MP

Sparse Subspace Clustering-Orthogonal Matching Pursuit (SSC-OMP) and SSC-Matching Pursuit (SSC-MP) are sparsity-based subspace clustering algorithms (see ref. 1, 2, 3 for SSC-OMP and ref. 3 for SSC-MP for more information).

Type

>> python3 ssc_mps.py

to run a simple face clustering experiment.

Tested with Python 3.4.2, numpy 1.11.0, scikit-learn 0.18.1.

References

1. E. L. Dyer, A. C. Sankaranarayanan, and R. G. Baraniuk, “Greedy feature selection for subspace clustering,” Journal of Mach. Learn. Research, vol. 14, pp. 2487–2517, 2013.
2. C. You, D Robinson, and R. Vidal, “Scalable sparse subspace clustering by orthogonal matching pursuit,” in IEEE Conf. on Comp. Vision and Pattern Recogn., 2016, pp. 3918– 3927.
3. M. Tschannen and H. Bölcskei, "Noisy Subspace Clustering via Matching Pursuits", arXiv:1612.03450, preprint, 2016.