AI Paper Atlas

Research map / Video / Video generation

CustomX: Unified Character, Action, and Scene Customization in Video World Models

Yitong Wang, Fangyun Wei, Hongyang Zhang, Bo Dai, Yan Lu

ECCV 2026 · Video › Video generation · approach: World-model video generation · 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 back view to the front-view character condition (ablation)CustomX · Character consistencyDINOv2 score: 0.556 -> 0.613Improved
add back view to the front-view character condition (ablation)CustomX · Character consistencyCLIP score: 0.628 -> 0.678Improved
add left and right views to front-back character condition (ablation)CustomX · Character consistencyDINOv2 score: 0.613 -> 0.698Improved
add left and right views to front-back character condition (ablation)CustomX · Character consistencyCLIP score: 0.678 -> 0.721Improved
add character anchor (first-frame only) (ablation)CustomX · Character consistencyDINOv2 score: 0.477 -> 0.597Improved
propagate anchor to per-frame anchors (ablation)CustomX · Character consistencyDINOv2 score: 0.597 -> 0.698Improved
propagate anchor to per-frame anchors (ablation)CustomX · Character consistencyCLIP score: 0.645 -> 0.721Improved
add per-frame anchor versus no anchor (ablation)CustomX · Character consistencyCLIP score: 0.529 -> 0.721Improved
add real-world videos to GTA-V training data (ablation)CustomX · Character consistency (real-world characters)DINOv2 score: 0.686 -> 0.755Improved
add real-world videos to GTA-V training data (ablation)CustomX · Character consistency (real-world characters)CLIP score: 0.718 -> 0.729Improved
distill 30-step model to 4 steps with DMD2CustomX · Character consistencyDINOv2 score: 0.698 -> 0.669Hurt
distill 30-step model to 4 steps with DMD2CustomX · Aesthetic qualityCLIP-Aesthetic: 5.665 -> 5.583Hurt

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