Research map / Video / Video generation
Cross-Modal Emotion Transfer for Emotion Editing in Talking Face Video
Chanhyuk Choi, Taesoo Kim, Donggyu Lee, Siyeol Jung, Taehwan Kim
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 contrastive loss L_cnt (ablation) | C-MET (L_recon only) · MEAD | Acc_emo: 49.43 -> 53.46 | Improved |
| add direction loss L_dir (ablation) | C-MET (L_recon + L_cnt) · MEAD | Acc_emo: 53.46 -> 55.91 | Improved |
| add contrastive loss L_cnt (ablation) | C-MET (L_recon only) · MEAD | FID: 88.951 -> 88.082 | Improved |
| add direction loss L_dir (ablation) | C-MET (L_recon + L_cnt) · MEAD | FID: 88.082 -> 90.804 | Hurt |
| use emotion2vec+large instead of Qwen2.5-Omni audio encoder (ablation) | C-MET · MEAD | Acc_emo: 52.06 -> 55.91 | Improved |
| use emotion2vec+large instead of Qwen2.5-Omni audio encoder (ablation) | C-MET · MEAD | AITV: 3.358 -> 2.643 | Improved |
| add contrastive loss L_cnt (ablation) | C-MET with Qwen2.5-Omni · MEAD | Acc_emo: 51.18 -> 52.06 | Improved |
| integrate C-MET into PD-FGC | PD-FGC · MEAD | Acc_emo: 33.36 -> 36.82 | Improved |
| integrate C-MET into PD-FGC | PD-FGC · MEAD | AITV: 1.247 -> 1.180 | Improved |
| integrate C-MET into EDTalk | EDTalk · MEAD | Acc_emo: 41.99 -> 55.91 | Improved |
| integrate C-MET into EDTalk | EDTalk · MEAD | FID: 76.423 -> 90.804 | Hurt |
| use C-MET (audio) instead of FLOAT (audio) | emotion editing · MEAD | Acc_emo: 13.21 -> 55.91 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- EDTalk — EDTalk: Efficient Disentanglement for Emotional Talking Head Synthesis (ECCV 2024) · better on 3 of 6 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
- EmotiveTalk: Expressive Talking Head Generation through Audio Information Decoupling and Emotional Video Diffusion (CVPR 2025 · similarity 0.89)
- Style2Talker: High-Resolution Talking Head Generation with Emotion Style and Art Style (AAAI 2024 · similarity 0.89)
- Emotional Face-to-Speech (ICML 2025 · similarity 0.89)
- DEEPTalk: Dynamic Emotion Embedding for Probabilistic Speech-Driven 3D Face Animation (AAAI 2025 · similarity 0.89)
- FlowVQTalker: High-Quality Emotional Talking Face Generation through Normalizing Flow and Quantization (CVPR 2024 · similarity 0.88)
- FG-EmoTalk: Talking Head Video Generation with Fine-Grained Controllable Facial Expressions (AAAI 2024 · similarity 0.88)
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.
