This is a Python implementation of the Principal Component Pursuit algorithm for robust PCA.
This implementation uses the fbpca implementation of approximate partial SVD for speed so you'll need to install that first.
TODO
Applied to the 'Escalator' dataset (using the code
in the demo.py script, this algorithm produces a video with frames that
look like:
Copyright 2015 Daniel Foreman-Mackey
This is open source software written by Dan Foreman-Mackey and released under the terms of the MIT license (see LICENSE).