AI Paper Atlas

Research map / Video / Video generation

MEVG : Multi-event Video Generation with Text-to-Video Models

Gyeongrok Oh, Jaehwan Jeong, Sieun Kim, Wonmin Byeon, Jinkyu Kim, Sungwoong Kim, Sangpil Kim

ECCV 2024 · Video › Video generation · approach: Compositional text-to-video diffusion · 24 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 structure-guided sampling on top of LFAI and dynamic noise (ablation)MEVG · Ablation user studyPreference (1-5): 2.78 -> 3.27Improved
add structure-guided sampling on top of LFAI and dynamic noise (ablation)MEVG · Ablation user studySemantics (1-5): 3.58 -> 3.47Hurt
add dynamic noise alone to base inversion (ablation)MEVG · Ablation user studyPreference (1-5): 2.89 -> 2.51Hurt
add last frame-aware inversion to dynamic noise (ablation)MEVG · Ablation user studyPreference (1-5): 2.51 -> 2.78Improved
add all three modules (LFAI, DN, SGS) to base DDIM inversion (ablation)MEVG · Ablation user studyTemporal consistency (1-5): 3.42 -> 3.70Improved
use a constant kappa instead of the decreasing noise schedule (ablation)Dynamic noise · Qualitative ablationvisual qualityHurt
apply last-frame guidance to more than the first frame (ablation)Last frame-aware inversion · Qualitative ablationmotion diversityHurt
raise SGS guidance scale from 7 to 50 (ablation)Structure-guided sampling · Hyper-parameter analysisCLIP-Text / CLIP-Image: CLIP-Text 30.674 -> 30.102; CLIP-Image 0.945 -> 0.954Mixed
MEVG vs Text2Video-Zero on CLIP-TextMEVG · Multi-event video generationCLIP-Text: 0.322 -> 0.309Hurt
MEVG vs Gen-L-Video and VidRD on human evaluationMEVG · Multi-event video generation, human evaluationTemporal consistency (1-5): 3.40 -> 3.82Improved

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