AI Paper Atlas

Research map / Video / Video generation

Consistent and Controllable Image Animation With Linear Motion Diffusion Transformers

Xin Ma, Yaohui Wang, Genyun Jia, Xinyuan Chen, Tien‐Tsin Wong, Cunjian Chen

TPAMI 2026 · Video › Video generation · approach: Efficient video diffusion models · 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 motion residual learning instead of direct frame prediction (ablation)MiraMo · VBench I2VI2V Average Consistency: 92.62 -> 94.02Improved
add motion residual learning instead of direct frame prediction (ablation)MiraMo · VBench I2VAverage Consistency: 93.62 -> 94.53Improved
add motion residual learning instead of direct frame prediction (ablation)MiraMo · VBench I2VDynamic Degree: 32.68 -> 35.12Improved
add DCTInit noise refinement on top of MRL (ablation)MiraMo · VBench I2VI2V Average Consistency: 94.02 -> 95.47Improved
add FFT noise refinement on top of MRL (ablation)MiraMo · VBench I2VI2V Average Consistency: 94.02 -> 94.93Improved
use DCT instead of FFT for noise refinement (ablation)MiraMo · VBench I2VI2V Average Consistency: 94.93 -> 95.47Improved
use DCT instead of FFT for noise refinement (ablation)MiraMo · VBench I2VDynamic Degree: 33.90 -> 35.93Improved
add FFT noise refinement on top of MRL (ablation)MiraMo · VBench I2VDynamic Degree: 35.12 -> 33.90Hurt
use synthetic (Wan2.1/HunyuanVideo) videos instead of real data for post-training (ablation)MiraMo T2V · VBench T2VMultiple Object: 52.91 -> 45.50Hurt
use synthetic videos instead of real data for post-training (ablation)MiraMo T2V · VBench T2Vaverage VBench score: over a 3% improvementImproved
use dependent residuals z_i - z_{i-1} instead of residuals to first frame (ablation)MiraMo · VBench I2Vvideo qualityHurt
use frame-wise temporal attention instead of full 3D attention over video tokens (ablation)MiraMo T2V · T2Vtemporal consistencyImproved

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