Research map / Video / Video generation
Responsive Listening Head Generation: A Benchmark Dataset and Baseline
Mohan Zhou, Yalong Bai, Wei Zhang, Ting Yao, Tiejun Zhao, Tao Mei
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 speaker visual input to audio-only listener model (ablation) | listener generator · ViCo D_test | FD angle: 8.69 -> 7.79 | Improved |
| add speaker audio input to visual-only listener model (ablation) | listener generator · ViCo D_test | FD expression: 18.85 -> 15.04 | Improved |
| add speaker audio input to visual-only listener model (ablation) | listener generator · ViCo D_test | FD translation: 7.54 -> 6.52 | Improved |
| add speaker audio and visual input versus mirroring the speaker (ablation) | listener generator · ViCo D_test | FD expression: 26.48 -> 15.04 | Improved |
| use correct audio instead of wrong audio (ablation) | listener generator · ViCo D_test | FD expression: 18.02 -> 15.04 | Improved |
| use auto-regressive LSTM decoder instead of parallel non-sequential layers (ablation) | listener generator · ViCo D_test | FD expression: 18.50 -> 15.04 | Improved |
| use auto-regressive LSTM decoder instead of parallel non-sequential layers (ablation) | listener generator · ViCo D_test | FD translation: 7.06 -> 6.52 | Improved |
| use learned responsive model instead of random perturbation | listener generator · ViCo D_test | FD expression: 44.67 -> 15.04 | Improved |
| use learned model instead of repeating sampled motion patterns (Simulation) | listener generator · ViCo D_ood | FD expression: 28.66 -> 22.83 | Improved |
| condition simulated motion patterns on the matching attitude (Simulation*) | Simulation baseline · ViCo D_test | FD angle: 10.81 -> 11.24 | No clear effect |
| audio input gives expression, visual input gives head motion (ablation) | listener generator · ViCo D_test | FD expression vs angle: 17.19 -> 18.85 | Mixed |
Closest papers by meaning
- GLARE: Generating Listening Heads with Appropriate REactions (NeurIPS 2026 · similarity 0.90)
- CustomListener: Text-guided Responsive Interaction for User-friendly Listening Head Generation (CVPR 2024 · similarity 0.88)
- EmotiveTalk: Expressive Talking Head Generation through Audio Information Decoupling and Emotional Video Diffusion (CVPR 2025 · similarity 0.87)
- Diffusion-based Realistic Listening Head Generation via Hybrid Motion Modeling (CVPR 2025 · similarity 0.87)
- LLM-driven Multimodal and Multi-Identity Listening Head Generation (CVPR 2025 · similarity 0.87)
- DualTalk: Dual-Speaker Interaction for 3D Talking Head Conversations (CVPR 2025 · 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.
