This is a collection of codes in C or in MATLAB to reproduce some of the results that are described in the paper: Frank Ong and Michael Lustig "Beyond Low Rank + Sparse: Multi-scale Low Rank Matrix Decomposition" https://arxiv.org/abs/1507.08751
You are encouraged to modify or distribute this code in any way you want. However, please acknowledge this code and cite the papers appropriately. For any questions about the code, please contact me (Frank Ong) at: frankong@berkeley.edu
##MATLAB Demos:
demo_hanning_decom: Perform multi-scale Low Rank Decomposition on a synthetic matrix with hanning matrices
demo_face_decom: Perform multi-scale low rank decomposition on a face image dataset
demo_dce_mri_decom: Perform multi-scale low rank decomposition on a fully sampled Dynamic Contrast Enhanced image dataset
Demos of the multi-scale low rank applied on surveillance videos and rating matrices are absent in the Matlab package as they are simply too slow to run on MATLAB.
##C Demos:
The C demo uses the Berkeley Advanced Reconstruction Toolbox (included as a submodule)
demo_dce_mri_decom: Perform multi-scale low rank decomposition on a fully sampled Dynamic Contrast Enhanced image dataset
demo_face_decom: Perform multi-scale low rank decomposition on a face image dataset
demo_hall_decom: Perform multi-scale low rank decomposition on surveillance video dataset