AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

Self-supervised Outdoor Scene Relighting

Ye Yu, Abhimitra Meka, Mohamed Elgharib, Hans-Peter Seidel, Christian Theobalt, William A. P. Smith

ECCV 2020 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Pretext-task visual pre-training · 53 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 the proposed neural-rendering relighting instead of Barron and Malikrelighting pipeline · BigTime time-lapse (15 sequences)l1 error: 0.115 -> 0.103Improved
use the proposed method instead of Barron and Malikrelighting pipeline · BigTime time-lapse (15 sequences)MSE: 0.24 -> 0.021Improved
use the proposed method instead of Barron and Malikrelighting pipeline · BigTime time-lapse (15 sequences)SSIM: 0.770 -> 0.760Hurt
use the proposed method instead of Yu and Smithrelighting pipeline · Relighting benchmark (original lighting 2)SSIM: 0.780 -> 0.850Improved
use the proposed method instead of Philip et al. (multi-view, sun model)relighting pipeline · Relighting benchmark (original lighting 6)l1 colour error: 0.083 -> 0.073Improved
use the proposed method instead of Barron and Malikrelighting pipeline · Relighting benchmark (original lighting 1)SSIM: 0.871 -> 0.879Hurt
add shadow channel s(p) to inverse rendering (ablation)InverseRenderNet · qualitative MegaDepthalbedo shadow bakingImproved
add cycle consistency loss to direct supervision and adversarial loss (ablation)neural renderer · MegaDepth trainingstability under smooth illumination changeImproved
compute cross projection loss at full resolution instead of downscaling by 4 (ablation)neural renderer · MegaDepth trainingsharpnessHurt
add self-reconstruction-only pre-training phase before cycle and adversarial losses (ablation)neural renderer · MegaDepth trainingconvergence speed and detailImproved

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