Research map / Video / Video generation
SpeakerVid-5M: A Large-Scale High-Quality Dataset for Audio-Visual Dyadic Interactive Human Generation
Youliang Zhang, Zhaoyang Li, Duomin Wang, jiahe zhang, Deyu Zhou, Zixin Yin, Xili Dai, Gang YU, Xiu Li
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add spatial set-by-set transformer (ablation) | AR audio-visual baseline · VidChatBench (Dyadic) | FID: 49.86 -> 35.67 | Improved |
| add spatial set-by-set transformer (ablation) | AR audio-visual baseline · VidChatBench (Conditioned) | FVD: 55.16 -> 34.64 | Improved |
| add spatial set-by-set transformer (ablation) | AR audio-visual baseline · VidChatBench (Dyadic) | ArcFace: 0.635 -> 0.749 | Improved |
| add training noise injection (ablation) | AR audio-visual baseline · VidChatBench (Dyadic) | FID: 35.67 -> 32.35 | Improved |
| add training noise injection (ablation) | AR audio-visual baseline · VidChatBench (Conditioned) | FID: 38.53 -> 34.72 | Improved |
| add training noise injection (ablation) | AR audio-visual baseline · VidChatBench (Dyadic) | Sync_conf: 2.541 -> 2.698 | Improved |
| add joint audio generation (ablation) | AR visual baseline · VidChatBench (Conditioned) | FID: 56.82 -> 57.03 | No clear effect |
| add joint audio generation (ablation) | AR visual baseline · VidChatBench (Dyadic) | FVD: 47.23 -> 36.90 | Improved |
| use dyadic audiovisual input instead of text conditioning | full AR baseline · VidChatBench | FID: 34.72 -> 32.35 | Improved |
| use dyadic audiovisual input instead of text conditioning | full AR baseline · VidChatBench | ArcFace: 0.758 -> 0.772 | Improved |
Closest papers by meaning
- TalkCuts: A Large-Scale Dataset for Multi-Shot Human Speech Video Generation (NeurIPS 2025 · similarity 0.89)
- INFP: Audio-Driven Interactive Head Generation in Dyadic Conversations (CVPR 2025 · similarity 0.88)
- Let Them Talk: Audio-Driven Multi-Person Conversational Video Generation (NeurIPS 2025 · similarity 0.88)
- VASA-1: Lifelike Audio-Driven Talking Faces Generated in Real Time (NeurIPS 2024 · similarity 0.87)
- The Audio-Visual Conversational Graph: From an Egocentric-Exocentric Perspective (CVPR 2024 · similarity 0.86)
- Friends-MMC: A Dataset for Multi-modal Multi-party Conversation Understanding (AAAI 2025 · similarity 0.86)
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.
