AI Paper Atlas

Research map / 3D vision / Depth and stereo

DuRP: Dual-Stage Physics-Embedded Learning for Joint Radiance and Polarization Restoration

Zhenshuo Yang, Qian He, Zhiyuan Liu, Baojie Fan, Jiandong Tian

ICML 2026 · 3D vision › Depth and stereo · approach: Neural light field and illumination imaging

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 the generalized polarization model instead of the traditional model (ablation)DuRP · SRPG test setPSNR DoLP: 25.594 -> 28.803Improved
add the generalized polarization model instead of the traditional model (ablation)DuRP · SRPG test setPSNR S0: 28.244 -> 33.696Improved
add the generalized polarization model instead of the traditional model (ablation)DuRP · SRPG test setPSNR AoLP: 12.543 -> 19.431Improved
embed the physical model into the network instead of direct polarization prediction (ablation)DuRP · SRPG test setPSNR DoLP: 24.355 -> 28.803Improved
embed the physical model into the network instead of direct polarization prediction (ablation)DuRP · SRPG test setPSNR AoLP: 16.242 -> 19.431Improved
add the refinement network (ablation)DuRP · SRPG test setPSNR S0: 31.652 -> 33.696Improved
add the refinement network (ablation)DuRP · SRPG test setPSNR AoLP: 18.662 -> 19.431Improved
add the polarization physical model in the P-Stage (ablation)DuRP · SRPG test setPSNR DoLP: DoLP PSNR decreasing by 3.16 dBImproved
add the polarization physical model in the P-Stage (ablation)DuRP · SRPG test setPSNR AoLP: AoLP PSNR decreasing by 2.65 dBImproved
add physical models in both stages versus fully direct prediction (ablation)DuRP · SRPG test setPSNR DoLP: DoLP and AoLP PSNR dropping by 4.44 dB and 3.19 dBImproved
use lightweight U-Net instead of Swin-T for AL-Net (ablation)DuRP · SRPG test setPSNR S0: 31.701 -> 33.700Improved
use FP16 AMP inferenceDuRP · SRPG test setFPS: 16.8 -> 26.4No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

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