AI Paper Atlas

Research map / 3D vision / 3D generation

Refining Few-Step Text-to-Multiview Diffusion via Reinforcement Learning

Ziyi Zhang, Li Shen, Deheng Ye, Yong Luo, Huangxuan Zhao, Meng Liu, Wei Yu, Lefei Zhang

CVPR 2026 · 3D vision › 3D generation · approach: Text-to-3D scene generation · 2 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 advantage learning with zigzag sampling (MV-ZigAL) versus MV-RDLMV-Adapter (LCM-SDXL) · Training promptsHyperScore overall: 9.03 -> 9.17Improved
add advantage learning with zigzag sampling (MV-ZigAL) versus MV-PGMV-Adapter (LCM-SDXL) · Training promptsHyperScore overall: 8.39 -> 9.17Improved
use advantage learning instead of standard policy gradient with zigzag pass (ablation)MVC-ZigAL · MATE-3DHyperScore overall: 6.86 -> 7.04Improved
use first-step zigzag instead of full-step zigzag (ablation)MVC-ZigAL · MATE-3DHyperScore overall: 6.91 -> 7.04Improved
use multiview-constrained optimization instead of weighted sum (w_mv=0.1) (ablation)MVC-ZigAL · MATE-3DHyperScore overall: 6.25 -> 7.04Improved
use multiview-constrained optimization instead of weighted sum (w_mv=0.5) (ablation)MVC-ZigAL · MATE-3DHyperScore overall: 6.83 -> 7.04Improved
use multiview-constrained optimization instead of weighted sum (w_mv=1.0) (ablation)MVC-ZigAL · MATE-3DHyperScore overall: 6.68 -> 7.04Improved
use adaptive threshold instead of fixed thresholds (7.5 or 9.0) (ablation)MVC-ZigAL · training reward curvessingle-view and joint-view rewards: 7.5 and 9.0 fixedImproved
use adaptive step size instead of fixed step sizes (0.01 or 0.1) (ablation)MVC-ZigAL · training reward curvesconstraint satisfaction: 0.01 and 0.1 fixedImproved
scale RL finetuning to MATE-3D prompt setMVC-ZigAL · MATE-3DImageReward: -0.846 -> 0.865Improved
add single-view-only reward (ZigAL) or weighted sum (ablation)ZigAL / WS-ZigAL · Training promptsjoint vs per-view rewardHurt

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