AI Paper Atlas

Research map / Video / Video generation

DVG-WM: Disentangled Video Generation Enables Efficient Embodied World Model for Robotic Manipulation

Ziyu Shan, Zhenyu Wu, Xiaofeng Wang, Zheng Zhu, Ziwei Wang

ECCV 2026 · Video › Video generation · approach: World-model video generation · 3 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 latent degradation instead of pixel degradation (ablation)DVG-WM refinement · LIBEROSSIM: 0.768 -> 0.783Improved
add latent degradation instead of pixel degradation (ablation)DVG-WM refinement · LIBEROFVD: 168.54 -> 152.36Improved
use flow matching from low-res latents instead of naive cascading conditioning (ablation)DVG-WM · LIBEROFVD: 187.69 -> 152.36Improved
use 256x384 preview resolution instead of 360p (ablation)DVG-WM preview stage · LIBEROPSNR: 18.812 -> 20.019Improved
add preview stage to refine-only model (ablation)DVG-WM · LIBEROFVD: 169.29 -> 152.36Improved
add refinement stage to preview-only model (ablation)DVG-WM · LIBEROLPIPS: 0.341 -> 0.120Improved
increase refinement steps K from 1 to 4 (ablation)DVG-WM · LIBEROFVD: 245.83 -> 152.36Improved
increase refinement steps K from 4 to 8 (ablation)DVG-WM · LIBEROFVD: 152.36 -> 150.42No clear effect
use disentangled two-stage DVG-WM instead of CogVideoX-5Bworld model · LIBEROFVD: 171.24 -> 152.36Improved
use DVG-WM instead of LVP-14B for speedworld model · LIBEROinference time: 354.2s -> 88.7sImproved
use DVG-WM instead of LVP-14Bworld model · LIBEROobject-level accuracy: 80% -> 89%Improved
use DVG-WM instead of CogVideoX with action expertreal-world manipulation · 9 real-world taskssuccess rate: 34.0 -> 62.2Improved

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