/pyMSLRDecomp

project repo for multi-scale low rank + sparse decomposition

Primary LanguageJupyter Notebook

Multi-scale Low-Rank Decomposition in Python

This project contains Python code to run multi-scale + low-rank matrix decomposition as introduced in this paper. The code in this repository is a port of the MATLAB implementation from the paper author(s).

Dependencies

At the time of writing, this code requires the following dependencies. (The code may or may not run with less or more recent versions.)

  1. Python 3.6
  2. Numpy 1.18.1
  3. Matplotlib 3.1.2
  4. skimage 0.16.2
  5. opencv-python 4.2.0

About this Project

Although this Python code took hours of effort to complete, primary credit belongs with the author(s) of the original MATLAB code. This work was motivated by my course project for ELEN E6876: Sparse and Low Rank Models for High Dimensional Data, at Columbia University in the spring of 2020. The course is taught by Prof. John Wright.

Project Authors

Dwiref Oza

Abhishek Bhan

References

These are included as links in the sections above, but listed here for clarity.

  1. F. Ong, M. Lustig, "Beyond Low Rank + Sparse: Multi-scale Low Rank Matrix Decomposition", arXiv
  2. MATLAB (And C implementation) by Frank Ong et. al github