AI Paper Atlas

Research map / Video / Video generation

MusicInfuser: Making Video Diffusion Listen and Dance

Susung Hong, Ira Kemelmacher-Shlizerman, Brian Curless, Steven M. Seitz

CVPR 2026 · Video › Video generation · approach: Efficient video diffusion models

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
select cross-attention layers by layer adaptability instead of all layers (ablation)MusicInfuser · Qwen3-Omni judgeOverall: 7.80 -> 8.14Improved
select layers by adaptability instead of evenly distributed layers (ablation)MusicInfuser · Qwen3-Omni judgeOverall: 7.62 -> 8.14Improved
select layers by adaptability instead of first layers (ablation)MusicInfuser · Qwen3-Omni judgeOverall: 7.99 -> 8.14Improved
use zero-initialized cross-attention (ZICA) instead of direct feature addition (ablation)MusicInfuser · Qwen3-Omni judgeOverall: 7.92 -> 8.14Improved
zero-initialize cross-attention output projection (ablation)MusicInfuser · Qwen3-Omni judgeVideo Quality Average: 7.82 -> 8.02Improved
add Beta-Uniform noise scheduling (ablation)MusicInfuser · Qwen3-Omni judgeOverall: 7.89 -> 8.14Improved
add in-the-wild dance data to AIST training (ablation)MusicInfuser · Qwen3-Omni judgeOverall: 7.95 -> 8.14Improved
add in-the-wild dance data to AIST training (ablation)MusicInfuser · Prompt alignmentPrompt Align Average: 7.96 -> 8.96Improved
randomly replace some detailed captions with base prompt (vs never or always) (ablation)MusicInfuser · Prompt alignmentPrompt Align Average: 8.58 -> 8.96Improved
use higher LoRA rank (ablation)MusicInfuser · VideoLLaMA 2 judgeOverall: 8.21 -> 8.33Improved
add LoRA adaptation (ablation)MusicInfuser · VideoLLaMA 2 judgeOverall: 8.18 -> 8.33Improved
adapt Mochi with music (MusicInfuser) vs base MochiMochi · Prompt alignmentStyle Capture: 7.98 -> 7.80Mixed

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