Learning Normalized Energy Models for Linear Inverse Problems
About
Generative diffusion models can provide powerful prior probability models for inverse problems in imaging, but existing implementations suffer from two key limitations: $(i)$ the prior density is represented implicitly, and $(ii)$ they rely on likelihood approximations that introduce sampling biases. We address these challenges by introducing a new energy-based model trained for denoising with a covariance-based regularization term that enforces consistency across different measurement conditions. The trained model can compute normalized posterior densities for diverse linear inverse problems, without additional retraining or fine tuning. In addition to preserving the sampling capabilities of diffusion models, this enables previously unavailable capabilities: energy-guided adaptive sampling that adjusts schedules on-the-fly, unbiased Metropolis-Hastings correction steps, and blind estimation of the degradation operator via Bayes rule. We validate the method on multiple datasets (ImageNet, CelebA, AFHQ) and tasks (inpainting, deblurring), demonstrating competitive or superior performance to established baselines.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Gaussian deblur | ImageNet 64 | LPIPS0.004 | 6 | |
| Half-mask inpainting | CelebA 64 x 64 | PSNR14 | 6 | |
| Inpainting | CelebA 64x64 | PSNR17.7 | 5 | |
| Gaussian Deblurring | CelebA 64x64 | PSNR32.27 | 5 | |
| Super-Resolution (x4) | CelebA 64 x 64 | PSNR22.41 | 3 | |
| Box Inpainting (30x30) | CelebA 64 x 64 | PSNR (dB)33.78 | 3 | |
| Inpainting (50x50 center-box) | CelebA 64 x 64 | PSNR (dB)26.45 | 3 | |
| Motion Deblurring | CelebA 64 x 64 | PSNR29.42 | 3 | |
| Random Inpainting (0.7%) | CelebA 64x64 | PSNR28.9 | 3 |