Research map / Video / Video generation
StyleTalk++: A Unified Framework for Controlling the Speaking Styles of Talking Heads
Suzhen Wang, Yifeng Ma, Yu Ding, Zhipeng Hu, Changjie Fan, Tangjie Lv, Zhidong Deng, Xin Yu
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 Transformer-XL with recurrence instead of standard transformer (teacher forcing) (ablation) | StyleTalk++ head pose · HeadMotion | SSIM: 0.566 -> 0.786 | Improved |
| use Transformer-XL instead of standard transformer with recurrence training (ablation) | StyleTalk++ head pose · HeadMotion | SSIM: 0.643 -> 0.786 | Improved |
| add spatial embedding transition (ablation) | StyleTalk++ head pose · HeadMotion | SSIM: 0.629 -> 0.786 | Improved |
| add head style discriminator (ablation) | StyleTalk++ head pose · HeadMotion | SSIM: 0.736 -> 0.786 | Improved |
| add head triplet loss (ablation) | StyleTalk++ head pose · HeadMotion | SSIM: 0.683 -> 0.786 | Improved |
| add temporal discriminator (ablation) | StyleTalk++ head pose · HeadMotion | SSIM: 0.710 -> 0.786 | Improved |
| use acoustic features instead of phonemes (ablation) | StyleTalk++ head pose · HeadMotion | SSIM: 0.593 -> 0.786 | Improved |
| use style code as decoder query (ablation) | StyleTalk++ expression · MEAD | M-LMD: 3.651 -> 3.249 | Improved |
| use adaptive (dynamic) feedforward layer (ablation) | StyleTalk++ expression · MEAD | M-LMD: 4.178 -> 3.249 | Improved |
| set K=8 instead of K=4 in dynamic feedforward (ablation) | StyleTalk++ expression · MEAD | M-LMD: 3.524 -> 3.249 | Improved |
| set K=8 instead of K=16 in dynamic feedforward (ablation) | StyleTalk++ expression · MEAD | M-LMD: 3.396 -> 3.249 | Improved |
| add expression style discriminator (ablation) | StyleTalk++ expression · MEAD | F-LMD: 2.483 -> 2.122 | Improved |
Closest papers by meaning
- Style2Talker: High-Resolution Talking Head Generation with Emotion Style and Art Style (AAAI 2024 · similarity 0.92)
- Say Anything with Any Style (AAAI 2024 · similarity 0.91)
- EmotiveTalk: Expressive Talking Head Generation through Audio Information Decoupling and Emotional Video Diffusion (CVPR 2025 · similarity 0.89)
- OmniTalker: One-shot Real-time Text-Driven Talking Audio-Video Generation With Multimodal Style Mimicking (NeurIPS 2025 · similarity 0.89)
- GoHD: Gaze-oriented and Highly Disentangled Portrait Animation with Rhythmic Poses and Realistic Expressions (AAAI 2025 · similarity 0.89)
- MimicTalk: Mimicking a personalized and expressive 3D talking face in minutes (NeurIPS 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.
