AI Paper Atlas

Research map / Image generation / Image restoration

StyleGAN knows Normal, Depth, Albedo, and More

Anand Bhattad, Daniel McKee, Derek Hoiem, David Forsyth

NeurIPS 2023 · Image generation › Image restoration · approach: Image harmonization and color transformation · 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
Use StyleGAN latent offset instead of supervised normal predictor (Omnidata-v1)StyleGAN · NormalsL1: 0.0501 -> 0.0834Hurt
Use StyleGAN latent offset instead of Omnidata-v1 on normalsStyleGAN · NormalsAngular error: 0.0750 -> 0.1216Hurt
Use StyleGAN latent offset instead of Omnidata-v2 for depthStyleGAN · DepthL1: 0.1237 -> 0.1019Improved
Use StyleGAN instead of X-task consistency for depthStyleGAN · DepthL1: 0.1390 -> 0.1019Improved
Use StyleGAN offset instead of DPT for segmentationStyleGAN · Segmentation meanmIoU: 87.7 -> 82.0Hurt
Use StyleGAN offset instead of DPT for segmentationStyleGAN · Segmentation meanAccuracy: 98.8 -> 98.1Hurt
Use StyleGAN offset instead of DPT for lamp segmentationStyleGAN · lampmIoU: 79.3 -> 71.9Hurt
Use StyleGAN offset instead of DPT for pillow segmentationStyleGAN · pillowmIoU: 82.5 -> 76.6Hurt
Use StyleGAN offset instead of DPT for painting segmentationStyleGAN · paintingmIoU: 92.9 -> 84.0Hurt
Use StyleGAN offsets rather than regression networks under relightingStyleGAN · Normals under 16 lighting conditionsangular change: 8 degreesImproved
Search for an offset that swaps left and right image halves (ablation)StyleGAN · Control taskqualitativeHurt
Use L1 loss versus task-specific losses (scale-invariant, classification) for offset search (ablation)StyleGAN · Depth/segmentationperformanceNo 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