AI Paper Atlas

Research map / 3D vision / Face and shape

Refine Now, Query Fast: A Decoupled Refinement Paradigm for Implicit Neural Fields

Tianyu Xiong, Skylar Wurster, Han Wei Shen

ICLR 2026 · 3D vision › Face and shape · approach: Implicit neural representation design · 1 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 decoupled refiner network on cached embeddings (DRR) to embedding-only model (ablation)DRR-Net · NyxPSNR (dB): 42.64 -> 44.69Improved
add decoupled refiner network to embedding-only model (ablation)DRR-Net · MPAS-OceanPSNR (dB): 45.22 -> 49.26Improved
add decoupled refiner network to embedding-only model (ablation)DRR-Net · Cloverleaf3DPSNR (dB): 45.44 -> 48.69Improved
add condition refiner on top of spatial refiner (ablation)DRR-Net · NyxPSNR (dB): 44.71 -> 44.69No clear effect
add condition refiner on top of spatial refiner (ablation)DRR-Net · MPAS-OceanPSNR (dB): 48.43 -> 49.26Improved
add Variational Pairs (VP-S) augmentation (ablation)DRR-Net · Cloverleaf3DPSNR (dB): 47.47 -> 49.26Improved
add Variational Pairs (VP-S) augmentation (ablation)DRR-Net · NyxPSNR (dB): 44.92 -> 44.61No clear effect
add Variational Coordinate (VC) augmentation (ablation)DRR-Net · NyxPSNR (dB): 44.92 -> 43.11Hurt
add Variational Pairs (VP-S) augmentation (ablation)FA-INR · Cloverleaf3DPSNR (dB): 32.40 -> 49.13Improved
add Variational Coordinate (VC) augmentation (ablation)Explorable-INR · Cloverleaf3DPSNR (dB): 45.03 -> 42.41Hurt
replace MLP-based FA-INR with DRR-Net on MPAS-Ocean (unstructured mesh)FA-INR · MPAS-OceanPSNR (dB): 52.13 -> 49.26Hurt
use DRR-Net instead of FA-INR for inference speedFA-INR · Nyxinference time (s): 287.2s -> 10.7sImproved

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