AI Paper Atlas

Research map / Video / Video generation

FLOAT: Generative Motion Latent Flow Matching for Audio-driven Talking Portrait

Taekyung Ki, Dongchan Min, Gyeongsu Chae

ICCV 2025 · Video › Video generation · approach: Talking face generation · 25 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 frame-wise AdaLN with masked self-attention instead of cross-attention (ablation)FLOAT · HDTFE-FID: 1.452 -> 1.229Improved
use frame-wise AdaLN instead of cross-attention (ablation)FLOAT · HDTFLSE-D: 7.757 -> 7.290Improved
use flow matching instead of epsilon-prediction diffusion (ablation)FLOAT · HDTFLSE-D: 9.922 -> 7.290Improved
use flow matching instead of x0-prediction diffusion (ablation)FLOAT · HDTFLSE-D: 9.048 -> 7.290Improved
use flow matching (10 NFE) instead of diffusion (50 steps) (ablation)FLOAT · HDTFFID: 21.190 -> 21.100No clear effect
add facial component perceptual loss to motion auto-encoder (ablation)motion auto-encoder · HDTFFID: 21.061 -> 19.803Improved
add facial component perceptual loss to motion auto-encoder (ablation)motion auto-encoder · RAVDESSFID: 28.866 -> 23.350Improved
increase NFE from 2 to 10 (ablation)FLOAT · HDTFFVD: 178.831 -> 162.052Improved
increase NFE from 10 to 20 (ablation)FLOAT · HDTFFVD: 162.052 -> 164.392No clear effect
increase audio guidance scale from 1 to 2 (ablation)FLOAT · RAVDESSLSE-D: 7.049 -> 6.994Improved
increase emotion guidance scale from 1 to 2 (audio scale 1) (ablation)FLOAT · RAVDESSE-FID: 1.555 -> 1.334Improved
add 3DMM head pose conditionFLOAT · HDTFFVD: 162.052 -> 126.663Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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