Research map / 3D vision / Gaussian splatting
EmoTaG: Emotion-Aware Talking Head Synthesis on Gaussian Splatting with Few-Shot Personalization
Haolan Xu, Keli Cheng, Lei Wang, Ning Bi, Xiaoming Liu
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 |
|---|---|---|---|
| remove Semantic Emotion Guidance (emotion distillation from DeepFace) (ablation) | EmoTaG · Emotional set (5s) | LMD: 3.067 -> 2.456 | Improved |
| add score distillation (L_Score) to the gate branch (ablation) | EmoTaG · Emotional set (5s) | Sync-C: 5.874 -> 6.147 | Improved |
| add KL emotion-distribution distillation (L_KL) to the residual branch (ablation) | EmoTaG · Emotional set (5s) | PSNR: 29.36 -> 29.95 | Improved |
| add gate branch to the motion decoder (ablation) | EmoTaG · Emotional set (5s) | Sync-C: 5.004 -> 6.147 | Improved |
| add residual branch to the motion decoder (ablation) | EmoTaG · Emotional set (5s) | LMD: 3.572 -> 2.456 | Improved |
| add AdaIN-based identity modulation (ablation) | EmoTaG · Emotional set (5s) | LMD: 4.021 -> 2.456 | Improved |
| use EmoTaG instead of InsTaG on the emotional test set | few-shot talking head · Emotional set (5s) | LMD: 3.428 -> 2.456 | Improved |
| use EmoTaG instead of Real3DPortrait on lip sync | few-shot talking head · Emotional set (5s) | Sync-C: 6.583 -> 6.147 | Hurt |
| use EmoTaG instead of InsTaG on OOD audio | few-shot talking head · Cross-identity audio | Sync-C: 4.722 -> 5.814 | Improved |
| use EmoTaG instead of InsTaG on rendering speed | few-shot talking head · Neutral set efficiency | FPS: 82.5 -> 76.4 | Hurt |
| use EmoTaG instead of TalkingGaussian on training time | talking head · Efficiency | training time: 27 min -> 11 min | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- InsTaG — InsTaG: Learning Personalized 3D Talking Head from Few-Second Video (CVPR 2025) · better on 13 of 14 shared results
- TalkingGaussian — TalkingGaussian: Structure-Persistent 3D Talking Head Synthesis via Gaussian Splatting (ECCV 2024) · better on 1 of 1 shared result
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
- EmotiveTalk: Expressive Talking Head Generation through Audio Information Decoupling and Emotional Video Diffusion (CVPR 2025 · similarity 0.90)
- DEEPTalk: Dynamic Emotion Embedding for Probabilistic Speech-Driven 3D Face Animation (AAAI 2025 · similarity 0.90)
- Style2Talker: High-Resolution Talking Head Generation with Emotion Style and Art Style (AAAI 2024 · similarity 0.89)
- InsTaG: Learning Personalized 3D Talking Head from Few-Second Video (CVPR 2025 · similarity 0.89)
- GoHD: Gaze-oriented and Highly Disentangled Portrait Animation with Rhythmic Poses and Realistic Expressions (AAAI 2025 · similarity 0.89)
- FaceTalk: Audio-Driven Motion Diffusion for Neural Parametric Head Models (CVPR 2024 · 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.
