Form a sparse a low rank factorization of a dataset using samples from the data
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.


Read Me !!

1. Setup
- Download and install OMPBox from Rob Rubinstein:

- Download and install oASIS solver:

- Add all folders in oASIS, OMPbox, and SEED to your path

2. To get started
- To run SEED on a synthetic dataset consisting of a union of subspaces, run uos_demo.m
- To run SEED on a real dataset consisting of a collection of face images under different illumination conditions, run face_demo.m (This face data is taken from a subset of the YaleB face database,

3. References
The paper associated with SEED:
E.L. Dyer, T.A. Goldstein, R. Patel, K.P. Kording, R.G. Baraniuk,"Self-Expressive Decompositions for Matrix Approximation and Clustering",

The paper associated with our column sampling method oASIS:
R. Patel, T.A. Goldstein, E.L. Dyer, A. Mirhoseini, R.G. Baraniuk, "oASIS: Adaptive Column Sampling for Kernel Matrix Approximation",

** Code and examples for the sampling method (oASIS) used in SEED can be found at: