Deep Generative Prior

Deep generative priors leverage the power of trained generative models, such as diffusion probabilistic models, to improve image reconstruction and signal recovery in various applications. Current research focuses on integrating these priors into Bayesian inference frameworks, often using algorithms like split Gibbs sampling or Metropolis-adjusted Langevin algorithms, and exploring their effectiveness across diverse imaging modalities (MRI, astronomy, ptychography). This approach offers significant advantages over traditional methods by enabling more robust and accurate reconstructions, particularly in challenging scenarios with limited data or noise, leading to improved image quality and more reliable quantitative results in diverse scientific and engineering fields.

Papers

July 31, 2024

Robust Simultaneous Multislice MRI Reconstruction Using Deep Generative Priors
Shoujin Huang, Guanxiong Luo, Yuwan Wang, Kexin Yang, Lingyan Zhang, Jingzhe Liu, Hua Guo, Min Wang, Mengye Lyu
Magnetic Resonance Imaging Reconstruction Echo Planar Imaging Deep Generative Prior Multi Slice

August 4, 2023

Generative Priors for MRI Reconstruction Trained from Magnitude-Only Images Using Phase Augmentation
Guanxiong Luo, Xiaoqing Wang, Mortiz Blumenthal, Martin Schilling, Erik Hans Ulrich Rauf, Raviteja Kotikalapudi, Niels Focke, Martin Uecker
Quantitative MRI Parallel Imaging Generative Image Prior Deep Generative Prior

May 2, 2023

Generalizing Dataset Distillation via Deep Generative Prior
George Cazenavette, Tongzhou Wang, Antonio Torralba, Alexei A. Efros, Jun-Yan Zhu
Dataset Distillation Pre Trained Generative Model Deep Generative Prior Cross Architecture Generalization

April 21, 2023

Plug-and-Play split Gibbs sampler: embedding deep generative priors in Bayesian inference
Florentin Coeurdoux, Nicolas Dobigeon, Pierre Chainais
Bayesian Inference Posterior Inference Plug and Play Diffusion Based Generative Model Sampling Strategy Gibbs Sampler Deep Generative Prior

February 3, 2023

Learning the Night Sky with Deep Generative Priors
Fausto Navarro, Daniel Hall, Tamas Budavari, Yashil Sukurdeep
LeArning Abstract Denoising Process Image Deconvolution Multi Frame Night Sky High Image Quality Deep Generative Prior

May 5, 2022

Compressive Ptychography using Deep Image and Generative Priors
Semih Barutcu, Doğa Gürsoy, Aggelos K. Katsaggelos
Deep Image Prior Generative Neural Network Deep Image Generative Prior X Ray Ptychography Coherent Diffraction Imaging Deep Generative Prior

April 24, 2022

Signal Recovery with Non-Expansive Generative Network Priors
Jorio Cocola
Compressed Sensing Signal Recovery Generative Image Prior Deep Generative Prior

March 18, 2022

Bayesian Inversion for Nonlinear Imaging Models using Deep Generative Priors
Pakshal Bohra, Thanh-an Pham, Jonathan Dong, Michael Unser
Ill Posed Inverse Problem Phase Retrieval Bayesian Inversion Nonlinear Inverse Problem Probabilistic Reconstruction Equivariant Imaging Deep Generative Prior