AI Paper Atlas

Research map / Video / Video generation

FlexAM: Flexible Appearance-Motion Decomposition for Versatile Video Generation Control

Mingzhi Sheng, Zekai Gu, Peng Li, Cheng Lin, Hao-Xiang Guo, Yingcong Chen, Yuan Liu

ECCV 2026 · Video › Video generation · approach: Trajectory-guided motion control · 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
use FlexAM motion video (multi-frequency + depth encoding) instead of DaS 3D tracking videoMotion transfer · Motion transferCLIP text-video alignment: 32.14 -> 32.55Improved
use FlexAM motion video instead of DaS 3D tracking videoMotion transfer · Motion transferCLIP temporal consistency: 0.968 -> 0.976Improved
use FlexAM instead of VACEMotion transfer · Motion transferVBench++ subject/background consistency: 0.86 -> 0.99Improved
use FlexAM instead of VACEMotion transfer · Motion transferVBench++ aesthetic quality: 0.57 -> 0.64Improved
use FlexAM instead of Wan2.2 FunMotion transfer · Motion transferVBench++ imaging quality: 0.68 -> 0.68No clear effect
use FlexAM instead of VACEPartial editing · Partial editing user studypreference, lighting consistency (%): 56% vs. 25%Improved
use FlexAM instead of VACEPartial editing · Partial editing user studypreference, unedited region preservation (%): 51% vs. 28%Improved
use FlexAM 5B single-GPU instead of 14B VACEPartial editing · Partial editing efficiencyinference time (s): 348s to 91sImproved
use FlexAM instead of task-specific camera methodsI2V camera control · RealEstate10K (100 sequences)rotation errorImproved
use FlexAM instead of Wan2.2 Fun Control CameraI2V camera control · RealEstate10K (100 sequences)translation errorHurt
add density-flexible training (stochastic point downsampling) vs DaS (ablation)FlexAM · Sparse-motion video reconstructionPSNR, SSIM, FVDImproved
add multi-frequency positional encoding (ablation)FlexAM motion video · Motion aliasing qualitativemotion identityImproved

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