AI Paper Atlas

Research map / Image generation / Image restoration

DeFlow: Learning Complex Image Degradations From Unpaired Data With Conditional Flows

Valentin Wolf, Andreas Lugmayr, Martin Danelljan, Luc Van Gool, Radu Timofte

CVPR 2021 · Image generation › Image restoration · approach: Efficient restoration architectures · 51 citations (OpenAlex)

Check 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
Use DeFlow-generated degradations instead of prior degradation pipelinesESRGAN super-resolution training · AIM-RWSRLPIPS: 0.394 -> 0.349Improved
Use DeFlow-generated degradations instead of prior degradation pipelinesESRGAN super-resolution training · NTIRE-RWSRLPIPS: 0.226 -> 0.218Improved
Use DeFlow-generated degradations instead of prior degradation pipelinesESRGAN super-resolution training · DPED-RWSRBRISQUE: 21.49 -> 21.13Improved
Increase flow depth from K=4 to K=16 (ablation)DeFlow · AIM-RWSRLPIPS: 0.362 -> 0.349Improved
Increase flow depth from K=8 to K=16 (ablation)DeFlow · AIM-RWSRLPIPS: 0.355 -> 0.349Improved
Use 4× rather than 8× downsampling for h(x) (ablation)DeFlow · AIM-RWSRLPIPS: 0.349 -> 0.429Hurt
Use 16× rather than 8× downsampling for h(x) (ablation)DeFlow · AIM-RWSRLPIPS: 0.349 -> 0.352Hurt
Add domain-invariant conditional information h(x) rather than using h(x)=0 (ablation)DeFlow · AIM-RWSRLPIPS: 0.412 -> 0.349Improved
Learn a correlated rather than non-learned latent shift (ablation)DeFlow · AIM-RWSRLPIPS: 0.405 -> 0.349Improved
Allow correlated rather than uncorrelated learned latent shift (ablation)DeFlow · AIM-RWSRLPIPS: 0.359 -> 0.349Improved
Set the same standard-normal base distribution for both domains (u=0) (ablation)DeFlow · AIM-RWSRdownstream super-resolution performanceHurt
Use DeFlow instead of CycleGAN for degradation learningReal-world super-resolution · AIM-RWSR, NTIRE-RWSR, and DPED-RWSRperceived image qualityImproved

Later papers that compared with this method

3 later accepted papers report DeFlow as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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