Papers › Bayesian Image Reconstruction using Deep Generative Models

Bayesian Image Reconstruction using Deep Generative Models

8 Dec 2020arXiv:2012.04567archive 2025-07-28

Razvan V Marinescu, Daniel Moyer, Polina Golland

Machine learning models are commonly trained end-to-end and in a supervised setting, using paired (input, output) data. Examples include recent super-resolution methods that train on pairs of (low-resolution, high-resolution) images. However, these end-to-end approaches require re-training every time there is a distribution shift in the inputs (e.g., night images vs daylight) or relevant latent variables (e.g., camera blur or hand motion). In this work, we leverage state-of-the-art (SOTA) generative models (here StyleGAN2) for building powerful image priors, which enable application of Bayes' theorem for many downstream reconstruction tasks. Our method, Bayesian Reconstruction through Generative Models (BRGM), uses a single pre-trained generator model to solve different image restoration tasks, i.e., super-resolution and in-painting, by combining it with different forward corruption models. We keep the weights of the generator model fixed, and reconstruct the image by estimating the Bayesian maximum a-posteriori (MAP) estimate over the input latent vector that generated the reconstructed image. We further use variational inference to approximate the posterior distribution over the latent vectors, from which we sample multiple solutions. We demonstrate BRGM on three large and diverse datasets: (i) 60,000 images from the Flick Faces High Quality dataset (ii) 240,000 chest X-rays from MIMIC III and (iii) a combined collection of 5 brain MRI datasets with 7,329 scans. Across all three datasets and without any dataset-specific hyperparameter tuning, our simple approach yields performance competitive with current task-specific state-of-the-art methods on super-resolution and in-painting, while being more generalisable and without requiring any training. Our source code and pre-trained models are available online: https://razvanmarinescu.github.io/brgm/.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

razvanmarinescu/brgm officialmentioned on GitHubtf report
razvanmarinescu/brgm-pytorch officialmentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Image DenoisingImage InpaintingImage ReconstructionImage RestorationImage Super-ResolutionSuper-ResolutionVariational Inference

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Denoising FFHQ BRGM LPIPS 0.24 #1 of 1 Archive leaderboard report
Image Denoising FFHQ 64x64 - 4x upscaling BRGM LPIPS 0.24 #1 of 1 Archive leaderboard report
Image Inpainting FFHQ 1024 x 1024 BRGM LPIPS 0.19 #1 of 2 Archive leaderboard report
Image Inpainting FFHQ 1024 x 1024 BRGM PSNR 21.33 #1 of 2 Archive leaderboard report
Image Inpainting FFHQ 1024 x 1024 BRGM RMSE 24.28 #1 of 2 Archive leaderboard report
Image Inpainting FFHQ 1024 x 1024 BRGM SSIM 0.84 #1 of 2 Archive leaderboard report
Image Inpainting FFHQ 1024 x 1024 SN-PatchGAN LPIPS 0.24 #2 of 2 Archive leaderboard report
Image Inpainting FFHQ 1024 x 1024 SN-PatchGAN PSNR 19.67 #2 of 2 Archive leaderboard report
Image Inpainting FFHQ 1024 x 1024 SN-PatchGAN RMSE 30.75 #2 of 2 Archive leaderboard report
Image Inpainting FFHQ 1024 x 1024 SN-PatchGAN SSIM 0.82 #2 of 2 Archive leaderboard report
Image Super-Resolution FFHQ 256 x 256 - 4x upscaling BRGM PSNR 24.16 #10 of 11 Archive leaderboard report
Image Super-Resolution FFHQ 256 x 256 - 4x upscaling BRGM SSIM 0.70 #10 of 11 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

ConvolutionPath Length RegularizationR1 RegularizationStyleGAN2Variational InferenceWeight Demodulation

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections