Research map / Image generation / Editing and style transfer
High-Fidelity Generalized Emotional Talking Face Generation With Multi-Modal Emotion Space Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 emotion consistency loss (ablation) | EAC · MEAD basic styles | EFD: 0.096 -> 0.065 | Helped |
| add emotion consistency loss (ablation) | EAC · MEAD basic styles | LMD: 2.40 -> 2.31 | Helped |
| add emotion consistency loss (ablation) | EAC · MEAD basic styles | Sync: 3.53 -> 3.57 | Helped |
| replace one-hot emotion encoding with CLIP (ablation) | emotion encoding · MEAD basic styles | EFD: 0.070 -> 0.065 | Helped |
| replace GPT2 emotion encoding with CLIP (ablation) | emotion encoding · MEAD basic styles | EFD: 0.067 -> 0.065 | Helped |
| replace MLPs with a Transformer in EAC (ablation) | EAC · MEAD basic styles | EFD: 0.122 -> 0.065 | Helped |
| replace GRUs with a Transformer in EAC (ablation) | EAC · MEAD basic styles | EFD: 0.088 -> 0.065 | Helped |
| add residual flow refinement at each scale (ablation) | HEF · MEAD basic styles | EFD: 0.082 -> 0.065 | Helped |
| add hierarchical flow updates instead of fixed 64 × 64 flow (ablation) | HEF · MEAD basic styles | EFD: 0.076 -> 0.065 | Helped |
| use text rather than audio or image emotion conditioning | Ours · MEAD | EFD: 0.069 / 0.071 -> 0.065 | Helped |
| use text rather than audio or image emotion conditioning | Ours · MEAD | FID: 15.91 / 15.89 -> 15.90 | No clear effect |
Closest papers by meaning
- Style2Talker: High-Resolution Talking Head Generation with Emotion Style and Art Style (AAAI 2024 · similarity 0.90)
- FlowVQTalker: High-Quality Emotional Talking Face Generation through Normalizing Flow and Quantization (CVPR 2024 · similarity 0.90)
- Emotional Face-to-Speech (ICML 2025 · similarity 0.90)
- EmotiveTalk: Expressive Talking Head Generation through Audio Information Decoupling and Emotional Video Diffusion (CVPR 2025 · similarity 0.88)
- DEEPTalk: Dynamic Emotion Embedding for Probabilistic Speech-Driven 3D Face Animation (AAAI 2025 · similarity 0.88)
- Cafe-Talk: Generating 3D Talking Face Animation with Multimodal Coarse- and Fine-grained Control (ICLR 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.