AI Paper Atlas

Research map / Video / Video generation

AI killed the Video Star. Audio-Driven Diffusion Model for Expressive Talking Head Generation

Baptiste Chopin, Tashvik Dhamija, Pranav Balaji, Yaohui Wang, Antitza Dantcheva

IJCV 2026 · Video › Video generation · approach: Talking face generation · 3 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 text input to Wav2Vec audio features (ablation)Dimitra · VoxCeleb2F-LMD: 6.16 -> 5.79Improved
add phoneme input to Wav2Vec audio features (ablation)Dimitra · VoxCeleb2F-LMD: 6.16 -> 6.06Improved
use three separate models for lips, expression and head pose instead of a single model (ablation)Dimitra · VoxCeleb2F-LMD: 6.15 -> 5.72Improved
replace Wav2Vec+text+phoneme with WavLM features and new training scheme (Dimitra++) (ablation)Dimitra · VoxCeleb2F-LMD: 5.72 -> 5.68Improved
replace Wav2Vec+text+phoneme with WavLM (Dimitra++) (ablation)Dimitra · VoxCeleb2AHD: 16.50 -> 12.58 (GT 18.91)Mixed
add audio encoding before each transformer layer instead of cross attention (ablation)Dimitra++ · VoxCeleb2F-LMD: 6.23 -> 5.68Improved
add audio encoding before each layer instead of only once (ablation)Dimitra++ · VoxCeleb2F-LMD: 5.98 -> 5.68Improved
add audio encoding by addition instead of concatenation (ablation)Dimitra++ · VoxCeleb2F-LMD: 6.02 -> 5.68Improved
add conditioning on the first pose (ablation)Dimitra++ · VoxCeleb2F-LMD: 6.06 -> 5.68Improved
use 8 attention heads (default) instead of 4 (ablation)Dimitra++ · VoxCeleb2F-LMD: 5.95 -> 5.68Improved
use default heads instead of 16 attention heads (ablation)Dimitra++ · VoxCeleb2M-LMD: 4.35 -> 4.36No clear effect
Dimitra++ vs Dimitra on CelebV-HQDimitra · CelebV-HQFID: 67.48 -> 64.51Improved

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