Research map / 3D vision / Face and shape
KMTalk: Speech-Driven 3D Facial Animation with Key Motion Embedding
Zhihao Xu, Shengjie Gong, Jiapeng Tang, Lingyu Liang, Yining Huang, Haojie Li, Shuangping Huang
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 KMTalk instead of SelfTalk | speech-driven 3D facial animation · VOCA-Test | LVE (x10^-5 mm): 3.2238 -> 2.2639 | Improved |
| use KMTalk instead of SelfTalk | speech-driven 3D facial animation · BIWI-Test-A | FDD (x10^-5 mm): 3.5761 -> 2.5446 | Improved |
| use KMTalk instead of SelfTalk on VOCA-Test upper-face dynamics | speech-driven 3D facial animation · VOCA-Test | FDD (x10^-7 mm): 4.0912 -> 4.0594 | No clear effect |
| add phoneme-based localization of key motions instead of uniform sampling (ablation) | KMTalk · BIWI-Test-A | LVE (x10^-4 mm): 4.1648 -> 3.9654 | Improved |
| add key motion decoder instead of selecting key frames from an existing method's output (ablation) | KMTalk · BIWI-Test-A | FDD (x10^-5 mm): 2.9546 -> 2.5446 | Improved |
| add audio guidance in cross-modal motion completion (ablation) | KMTalk · BIWI-Test-A | LVE (x10^-4 mm): 4.8859 -> 3.9654 | Improved |
| swap the ASR used for phoneme extraction (Auto-avsr to Whisper) (ablation) | KMTalk · BIWI-Test-A | FDD (x10^-5 mm): 2.5446 -> 2.8141 (Whisper-large) | No clear effect |
| shift key motion indices by one frame (ablation) | KMTalk · BIWI-Test-A | FDD (x10^-5 mm): 2.5446 -> 2.6420 | No clear effect |
| apply the key-motion progressive learning scheme to existing methods | FaceFormer, CodeTalker, SelfTalk · BIWI-Test-A | FDD (x10^-5 mm): SelfTalk 3.5761 -> 2.8668 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CodeTalker — CodeTalker: Speech-Driven 3D Facial Animation With Discrete Motion Prior (CVPR 2023) · better on 3 of 3 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
- Probabilistic Speech-Driven 3D Facial Motion Synthesis: New Benchmarks Methods and Applications (CVPR 2024 · similarity 0.93)
- DEEPTalk: Dynamic Emotion Embedding for Probabilistic Speech-Driven 3D Face Animation (AAAI 2025 · similarity 0.92)
- Seeing Speech: Learning Visible Articulatory Dynamics for Speech-Driven 3D Facial Animation (NeurIPS 2026 · similarity 0.91)
- ProsodyTalker: 3D Visual Speech Animation via Prosody Decomposition (AAAI 2025 · similarity 0.90)
- Mimic: Speaking Style Disentanglement for Speech-Driven 3D Facial Animation (AAAI 2024 · similarity 0.90)
- KeyFace: Expressive Audio-Driven Facial Animation for Long Sequences via KeyFrame Interpolation (CVPR 2025 · similarity 0.90)
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.
