AI Paper Atlas

Research map / Video / Video generation

Make-An-Animation: Large-Scale Text-conditional 3D Human Motion Generation

Samaneh Azadi, Akbar Shah, Thomas Hayes, Devi Parikh, Sonal Gupta

ICCV 2023 · Video › Video generation · approach: Text-to-motion diffusion · 56 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
pre-train on 35M text pseudo-pose dataset before motion capture fine-tuning (ablation)Make-An-Animation · 400 in-the-wild prompts and HumanML3Dhuman preference and stabilityImproved
train on mocap data from scratch without pose pre-training (ablation)Make-An-Animation · mocap trainingtraining stabilityHurt
use U-Net architecture instead of decoder-only autoregressive transformer (ablation)Make-An-Animation · 400 AMT promptsrater preference: 58% preferred U-NetImproved
add temporal attention layers (ablation)Make-An-Animation · 400 AMT promptsrater preference: 54% preferred temporal attentionImproved
add unconditional training on GTA-Human data (ablation)Make-An-Animation · motion trainingconvergence and sample qualityNo clear effect
Make-An-Animation vs T2M on HumanML3D R-Precision without optimizing for the 263-dim representationMake-An-Animation · HumanML3D testR-Precision: 0.6755 vs 0.740 (T2M)Mixed

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