atlas

Research map / 3D vision / NeRF and radiance fields

GM-NeRF: Learning Generalizable Model-Based Neural Radiance Fields From Multi-View Images

CVPR 2023 · 3D vision › NeRF and radiance fields · approach: Sparse-input NeRF regularization · 30 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (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 multi-view appearance blending (ablation)Base GM-NeRF · Multi-GarmentPSNR: 25.22 -> 27.08Helped
add perceptual loss (ablation)Base GM-NeRF · Multi-GarmentPSNR: 25.22 -> 27.75Helped
add neural rendering alongside MAB and PL (ablation)GM-NeRF · Multi-GarmentPSNR: 28.72 -> 30.18Helped
use proposed visibility-based and geometry-guided attention instead of AvgPool (ablation)GM-NeRF · THuman2.0PSNR: 27.81 -> 28.88Helped
use proposed multi-view fusion instead of IBRAtt (ablation)GM-NeRF · THuman2.0SSIM: 0.9345 -> 0.9518Helped
use proposed multi-view fusion instead of AvgPool (ablation)GM-NeRF · THuman2.0LPIPS: 0.05059 -> 0.03349Helped
use GM-NeRF rather than the strongest listed generalizable baselineMulti-Garment unseen identity · Multi-GarmentPSNR: 28.61 -> 30.18Helped
use GM-NeRF rather than the strongest listed generalizable baselineTHuman2.0 unseen identity · THuman2.0LPIPS: 0.0734 -> 0.0335Helped
use GM-NeRF rather than the strongest listed generalizable baselineGeneBody unseen identity · GeneBodyPSNR: 24.71 -> 23.90Hurt
fine-tune GM-NeRF pretrained on Multi-GarmentGM-NeRF on ZJUMocap · ZJUMocap novel pose synthesisPSNR: 26.68 -> 27.63Helped
fine-tune GM-NeRF pretrained on Multi-GarmentGM-NeRF on ZJUMocap · ZJUMocap novel view synthesisLPIPS: 0.0904 -> 0.0733Helped
fine-tune GM-NeRF on training dataGM-NeRF geometry reconstruction · Multi-GarmentChamfer: 0.7175 -> 0.3721Helped

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