AI Paper Atlas

Research map / Video / Video generation

Conversational Human Audio-visual Talking Dialogue Generation

Junhao Song, Lluis Guasch, Xilin He, zhongyu yang, Yingfang Yuan, Weicheng Xie, Linlin Shen, Lin Haijun, Shizhe Liu, Wei Pang, Siyang Song

ECCV 2026 · Video › Video generation · approach: Talking face generation · 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 interactive audio refinement (IAR) (ablation)CHAT · CHAT evaluation set (1000 dialogues, first 10 s)FRCorr (x1e-2): 32.45 -> 50.02Improved
add emotion representations (ablation)CHAT · CHAT evaluation set (1000 dialogues, first 10 s)Emo-Acc: 38.3% -> 78.4%Improved
add interactive words (ablation)CHAT · CHAT evaluation set (1000 dialogues, first 10 s)FRDiv (x1e-2): 9.34 -> 15.34Improved
add sound environment (ablation)CHAT · CHAT evaluation set (1000 dialogues, first 10 s)FID: 18.45 -> 17.33Improved
add interactive facial behaviour refinement (IFBR) (ablation)CHAT · CHAT evaluation set (1000 dialogues, first 10 s)FID: 38.67 -> 17.33Improved
add responsive facial behaviour generation (RFBG) (ablation)CHAT · CHAT evaluation set (1000 dialogues, first 10 s)FRCorr (x1e-2): 25.78 -> 50.02Improved
add silent facial behaviour generation instead of interpolation (ablation)CHAT · CHAT evaluation set (1000 dialogues, first 10 s)FRCorr (x1e-2): 49.45 -> 50.02Improved
add temporal continuity refinement (TCR) (ablation)CHAT · CHAT evaluation set (1000 dialogues, first 10 s)FVD: 421.34 -> 365.03Improved
use CHAT instead of ReactDiff for reactionCHAT · CHAT evaluation set (1000 dialogues, first 10 s)FRCorr (x1e-2): 48.05 -> 50.02Improved
use CHAT instead of Hallo3CHAT · CHAT evaluation set (1000 dialogues, first 10 s)FVD: 362.23 vs 365.03Hurt
pre-train on CHAT-AVD-50k before fine-tuningPerFRDiff · REACT 2024 testFRCorr (x1e-2): 37.21 -> 40.11Improved
pre-train on CHAT-AVD-50k before fine-tuningReactDiff · REACT 2024 testFRDist: 86.70 -> 83.87Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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