DeepTensor: Low-Rank Tensor Decomposition with Deep Network Priors

Paper: https://arxiv.org/abs/2204.03145

Contents

• data/ -- contains mat files required to run some python scripts. The folder contains a README listing the sources
  • cat_video.mat -- required for run_figure11.py and run_figure11_TV.py
  • icvl.mat -- required for run_figure9.py and run_figure9_bm3d.py
  • pet.mat -- required for run_figure12.py and run_figure12_TV.py
  • weizzman.mat -- required for run_figure2.py
• modules/ -- contains several python scripts required for all experiments
  • deep_models/ -- folder downloaded from https://github.com/DmitryUlyanov/deep-image-prior
  • cosine_annealing_with_warmup.py -- code for cosine annealing scheduler downloaded from https://github.com/katsura-jp/pytorch-cosine-annealing-with-warmup
  • deep_decoder.py -- convenience code downloaded from https://github.com/reinhardh/supplement_deep_decoder
  • deep_prior.py -- convenience code downloaded from https://github.com/DmitryUlyanov/deep-image-prior
  • lin_inverse.py -- implements functions required for linear inverse problems
  • losses.py -- implements losses required for training include L1, L2, and 2D TV
  • models.py -- implements multilayer perceptron, and multi-dimensional (1D, 2D, and 3D) U-Net
  • spectral.py -- required for run_figure6.py
  • utils.py -- miscellaneous utilities
• requirements.txt -- contains a list of requirements to run the code base. Install the required packages using pip install -r requirements.txt
• run_figure2.py -- run this to replicate results in figure 2
• run_figure6.py -- run this to replicate results in figure 6
• run_figure7.py -- run this to replicate results in figure 7
• run_figure9.py -- run this to replicate results with DeepTensor and SVD in figure 9
• run_figure9_bm3d.py -- run this to replicate results with BM3D in figure 9
• run_figure11.py -- run this to replicate results with DeepTensor in figure 11
• run_figure11_TV.py -- run this to replicate results with TV in figure 11
• run_figure12.py -- run this to replicate results with DeepTensor in figure 12
• run_figure12_TV.py -- run this to replicate results with TV in figure 12
• run_table1.py -- run this to replicate results in table 1