AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

MAVFlow: Preserving Paralinguistic Elements with Conditional Flow Matching for Zero-Shot AV2AV Multilingual Translation

Sungwoo Cho, Jeongsoo Choi, Sungnyun Kim, Se-Young Yun

ICCV 2025 · Image generation › Autoregressive and latent generation · approach: Neural text-to-speech synthesis · 1 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 visual guidance to audio guidance (ablation)MAVFlow · CREMA-DEmo-Acc (%): 35.85 -> 36.46Improved
add visual guidance to audio guidance (ablation)MAVFlow · MuAViCspeaker similarity: 0.487 -> 0.504Improved
add visual guidance to audio guidance (ablation)MAVFlow · MuAViCMCD-DTW: 7.50 -> 7.37Improved
add audio guidance (x-vector) to no guidance (ablation)MAVFlow · CREMA-Dspeaker similarity: 0.167 -> 0.391Improved
add audio guidance to no guidance (ablation)MAVFlow · MuAViCspeaker similarity: 0.057 -> 0.487Improved
add visual guidance alone without audio (ablation)MAVFlow · CREMA-DEmo-Acc (%): 28.66 -> 26.83Hurt
add visual guidance alone without audio (ablation)MAVFlow · CREMA-Dspeaker similarity: 0.167 -> 0.174No clear effect
replace AV2AV renderer with CFM-based MAVFlowAV2AV · CREMA-DEmo-Acc (%): 33.66 -> 36.46Improved
replace AV2AV renderer with MAVFlowAV2AV · MuAViC Fr-Enspeaker similarity: 0.31 -> 0.51Improved
replace AV2AV renderer with MAVFlowAV2AV · MuAViC It-EnMCD-DTW-SL: 14.75 -> 11.43Improved
replace AV2AV renderer with MAVFlowAV2AV · MuAViC Es-EnASR-BLEU: 26.57 -> 26.97Improved
replace AV2AV renderer with MAVFlowAV2AV · MuAViCMOS naturalness: 3.58 -> 4.01Improved

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