AI Paper Atlas

Research map / 3D vision / 3D generation

Lyra: Generative 3D Scene Reconstruction via Video Diffusion Model Self-Distillation

Sherwin Bahmani, Tianchang Shen, Jiawei Ren, Jiahui Huang, Yifeng Jiang, Haithem Turki, Andrea Tagliasacchi, David B. Lindell, Zan Gojcic, Sanja Fidler, Huan Ling, Jun Gao and 1 more

ICLR 2026 · 3D vision › 3D generation · approach: Score distillation for text-to-3D · 31 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
train 3DGS decoder with self-distillation from video model instead of real multi-view data only (ablation)Lyra · Lyra dataset ablationPSNR: 19.08 -> 24.77Improved
add real multi-view data to self-distillation (ablation)Lyra · Lyra dataset ablationPSNR: 24.77 -> 24.74No clear effect
add depth supervision loss (ablation)Lyra · Lyra dataset ablationPSNR: 24.31 -> 24.77Improved
add opacity-based pruning (ablation)Lyra · Lyra dataset ablationPSNR: 24.55 -> 24.77Improved
add opacity-based pruning (ablation)Lyra · render time at 704x1280ms: 30ms -> 18msImproved
add LPIPS loss (ablation)Lyra · Lyra dataset ablationLPIPS: 0.370 -> 0.224Improved
learn multi-view fusion across trajectories instead of independent fusion (ablation)Lyra · Lyra dataset ablationPSNR: 17.73 -> 24.77Improved
use Transformer plus Mamba-2 blocks instead of Transformer only (ablation)Lyra · Lyra dataset ablationPSNR: 24.58 -> 24.77Improved
operate 3DGS decoder in video latent space instead of pixel space (ablation)Lyra · memoryout-of-memoryImproved
use Lyra instead of Bolt3DLyra · RealEstate10KPSNR: 21.54 -> 21.79Improved
use Lyra instead of WonderlandLyra · DL3DVPSNR: 16.64 -> 20.09Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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