Research map / Video / Video generation
DiTaiListener: Controllable High Fidelity Listener Video Generation with Diffusion
Maksim Siniukov, Di Chang, Minh Tran, Hongkun Gong, Ashutosh Chaubey, Mohammad Soleymani
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 |
|---|---|---|---|
| generate listener video directly in pixel space with video diffusion instead of 3DMM motion codes plus renderer | DiTaiListener · ViCo test FID | FID: 49.75 -> 10.14 | Improved |
| generate listener video directly in pixel space with video diffusion | DiTaiListener · RealTalk test FID | FID: 30.55 -> 7.99 | Improved |
| add CTM-Adapter (causal temporal attention) in place of non-causal spatial attention (ablation) | DiTaiListener · ViCo test FID | FID: 11.56 -> 10.14 | Improved |
| add CTM-Adapter (ablation) | DiTaiListener · RealTalk test FD expression (3DMM) | FD (expression): 15.10 -> 14.28 | Improved |
| add text control (ablation) | DiTaiListener · ViCo test FD expression (3DMM) | FD (expression): 18.88 -> 17.49 | Improved |
| add text control (ablation) | DiTaiListener · RealTalk test FD expression (3DMM) | FD (expression): 15.62 -> 14.28 | Improved |
| add text control (ablation) | DiTaiListener · ViCo test FVD | FVD: 57.71 -> 53.54 | Improved |
| add CTM-Adapter (ablation) | DiTaiListener · ViCo test PSNR/LPIPS | PSNR: 20.77 vs 20.77 | No clear effect |
| add CTM-Adapter (ablation) | DiTaiListener · ViCo test SID pose (3DMM) | SID (pose): 3.71 -> 3.74 | Hurt |
| use DiTaiListener-Edit frame smoothing for long videos instead of prompt travel or teacher forcing (ablation) | DiTaiListener long video generation · qualitative long video comparison | smoothness / transition artifacts | Improved |
| compare against 3DMM-based baselines in user study | DiTaiListener · ViCo user study | overall preference: 12.83% -> 64.03% | Improved |
Closest papers by meaning
- Diffusion-based Realistic Listening Head Generation via Hybrid Motion Modeling (CVPR 2025 · similarity 0.91)
- MoDiTalker: Motion-Disentangled Diffusion Model for High-Fidelity Talking Head Generation (AAAI 2025 · similarity 0.90)
- DAWN: Dynamic Frame Avatar with Non-autoregressive Diffusion Framework for Talking head Video Generation (ICLR 2025 · similarity 0.90)
- OmniTalker: One-shot Real-time Text-Driven Talking Audio-Video Generation With Multimodal Style Mimicking (NeurIPS 2025 · similarity 0.90)
- VLOGGER: Multimodal Diffusion for Embodied Avatar Synthesis (CVPR 2025 · similarity 0.89)
- Long-Term TalkingFace Generation via Motion-Prior Conditional Diffusion Model (ICML 2025 · similarity 0.89)
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.
