AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Diffusion Prior-Based Amortized Variational Inference for Noisy Inverse Problems

Sojin Lee, Dogyun Park, Inho Kong, Hyunwoo J. Kim

ECCV 2024 · Image generation › Fast diffusion sampling · approach: Diffusion posterior sampling · 6 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add integral KL divergence (IKL) loss (ablation)DAVI · FFHQ Gaussian deblurFID: 50.69 -> 37.61Improved
add Perturbed Posterior Bridge with no perturbation (h=0) (ablation)DAVI (IKL only) · FFHQ Gaussian deblurFID: 37.61 -> 49.37Hurt
add perturbation h=0.01 to the bridge (ablation)DAVI · FFHQ Gaussian deblurFID: 49.37 -> 47.41Improved
sample bridge index a from beta distribution instead of uniform (ablation)DAVI · FFHQ Gaussian deblurFID: 47.41 -> 34.34Improved
use monotonically increasing perturbation schedule instead of constant (ablation)DAVI · FFHQ Gaussian deblurFID: 34.34 -> 29.92Improved
single-step amortized inference instead of iterative samplinginverse problem solvers · FFHQ 4x super-resolutionsec/image: 1.7 -> 0.04 (vs DDRM)Improved
use DAVI instead of noise-scale-dependent solvers when sigma_y is unknownGaussian deblur · ImageNet 256FID: 157.38 -> 63.37 (vs DDRM)Improved
use DAVI instead of RED-diff on Gaussian deblurinverse problem solvers · FFHQ 256PSNR: 26.44 -> 25.46Mixed
use DAVI instead of DiffPIR on Gaussian deblurinverse problem solvers · ImageNet 256FID: 62.48 -> 63.29No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy