Recovery Analysis for Plug-and-Play Priors using the Restricted Eigenvalue Condition

Authors: Jiaming Liu, Salman Asif, Brendt Wohlberg, Ulugbek Kamilov

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We also present numerical results comparing the recovery performance of PnP/RED in compressive sensing against that of recent compressive sensing algorithms based on generative models.
Researcher Affiliation Academia Jiaming Liu Washington University in St. Louis jiaming.liu@wustl.edu M. Salman Asif University of California, Riverside sasif@ece.ucr.edu Brendt Wohlberg Los Alamos National Laboratory brendt@ieee.org Ulugbek S. Kamilov Washington University in St. Louis kamilov@wustl.edu
Pseudocode No The paper describes algorithms using mathematical equations (e.g., (5) and (6)) but does not provide structured pseudocode or algorithm blocks.
Open Source Code Yes The code for our numerical evaluation is available at: https://github.com/wustl-cig/pnp-recovery.
Open Datasets Yes The training data for the denoiser is generated by adding AWGN to the images from the BSD500 [78] and DIV2K datasets [79].
Dataset Splits No The paper mentions training and test datasets (e.g., BSD500, DIV2K, Celeb A HQ, FFHQ) and their use, but does not explicitly define or specify details about a validation dataset or its split.
Hardware Specification No The paper discusses training deep models and running experiments but does not provide specific hardware details such as GPU or CPU models.
Software Dependencies No The paper mentions software components like 'Dn CNN architecture' and 'spectral normalization' but does not provide specific version numbers for these or other dependencies (e.g., Python version, deep learning framework versions).
Experiment Setup Yes We pre-train several deep models as denoisers for σ [1, 15], using σ intervals of 0.5, and use the denoiser achieving the best PSNR value in each experiment.