Research map / Video / Video generation
Avatar Forcing: Real-Time Interactive Head Avatar Generation for Natural Conversation
Taekyung Ki, Sangwon Jang, Jaehyeong Jo, Jaehong Yoon, Sung Ju Hwang
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 |
|---|---|---|---|
| use 10 blocks instead of 5 (smaller latency, larger blocks count) (ablation) | Avatar Forcing · RealTalk | latency: 0.5 -> 0.3 (latency); FVD 170.87 -> 222.47 | Mixed |
| use 2 blocks instead of 5 (ablation) | Avatar Forcing · RealTalk | FVD / LSE-D / latency: FVD 170.87 -> 155.81; LSE-D 6.723 -> 6.555; latency 0.5 -> 1.5 | Mixed |
| add user motion input to the Dual Motion Encoder (ablation) | Avatar Forcing · RealTalk | Reactiveness and Motion Richness | Improved |
| add DPO preference fine-tuning with synthetic losing samples (ablation) | Avatar Forcing · RealTalk | rPCC-Exp, rPCC-Pose, SID, Var | Improved |
| add look-ahead to the blockwise causal mask (ablation) | Avatar Forcing · RealTalk | temporal smoothness (qualitative) | Improved |
| use diffusion forcing instead of autoregressive diffusion (ablation) | Avatar Forcing · long-horizon generation | motion drift | Improved |
| use causal blockwise diffusion forcing instead of bidirectional INFP* | Avatar Forcing · RealTalk | latency: 3.4 -> 0.5 | Improved |
Closest papers by meaning
- VASA-1: Lifelike Audio-Driven Talking Faces Generated in Real Time (NeurIPS 2024 · similarity 0.89)
- DAWN: Dynamic Frame Avatar with Non-autoregressive Diffusion Framework for Talking head Video Generation (ICLR 2025 · similarity 0.89)
- InstructAvatar: Text-Guided Emotion and Motion Control for Avatar Generation (AAAI 2025 · similarity 0.88)
- AvatarGO: Zero-shot 4D Human-Object Interaction Generation and Animation (ICLR 2025 · similarity 0.87)
- FlowVQTalker: High-Quality Emotional Talking Face Generation through Normalizing Flow and Quantization (CVPR 2024 · similarity 0.87)
- From Audio to Photoreal Embodiment: Synthesizing Humans in Conversations (CVPR 2024 · similarity 0.87)
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.
