Research map / Video / Video generation
Identity-Preserving Video Dubbing Using Motion Warping
Ruzhang Liu, Qinjie Lin, Yunfei Liu, Lijian Lin, Ye Zhu, Yu Li, Chuhua Xian, Fa-Ting Hong
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 the alignment module (AVAU) instead of a conv audio encoder (ablation) | IPTalker · HDTF | PSNR: 29.6988 -> 30.8796 | Improved |
| add the alignment module (AVAU) instead of a conv audio encoder (ablation) | IPTalker · HDTF | LSE-C: 7.9878 -> 8.2741 | Improved |
| add the SPADE decoder instead of a conv decoder (ablation) | IPTalker · HDTF | PSNR: 29.8337 -> 30.8796 | Improved |
| add the SPADE decoder instead of a conv decoder (ablation) | IPTalker · HDTF | LPIPS: 0.0626 -> 0.0582 | Improved |
| add the cross-modalities encoder E_cm (ablation) | IPTalker · HDTF | LSE-D: 7.2375 -> 7.0218 | Improved |
| add the cross-modalities encoder E_cm (ablation) | IPTalker · HDTF | LSE-C: 8.0697 -> 8.2741 | Improved |
| add the cross-modalities encoder E_cm (ablation) | IPTalker · HDTF | SSIM: 0.6265 -> 0.6365 | Improved |
| identity fine-tuning for 300 steps | IPTalker (identity fine-tuning) · HDTF | PSNR: 30.8919 -> 34.2662 | Improved |
| identity fine-tuning for 100 steps | IPTalker (identity fine-tuning) · HDTF | LPIPS: 0.0584 -> 0.0388 | Improved |
| increase fine-tuning from 200 to 300 steps | IPTalker (identity fine-tuning) · HDTF | LSE-D: 6.9742 -> 6.9888 | No clear effect |
| compare IPTalker with DINet (training on HDTF only) | IPTalker vs DINet · HDTF | PSNR: 27.7341 -> 30.8796 | Improved |
| compare IPTalker with Hallo on profile-heavy data | IPTalker vs Hallo · VFHQ | LSE-C: 6.5482 -> 6.0691 | Hurt |
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.
- DINet — DINet: Deformation Inpainting Network for Realistic Face Visually Dubbing on High Resolution Video (AAAI 2023) · better on 10 of 10 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
- From Inpainting to Editing: Unlocking Robust Mask-Free Visual Dubbing via Generative Bootstrapping (ICML 2026 · similarity 0.90)
- OmniSync: Towards Universal Lip Synchronization via Diffusion Transformers (NeurIPS 2025 · similarity 0.88)
- CoSyncDiT: Cognitive Synchronous Diffusion Transformer for Movie Dubbing (ECCV 2026 · similarity 0.88)
- DiVAS: Video and Audio Synchronization with Dynamic Frame Rates (CVPR 2024 · similarity 0.86)
- Synergizing Motion and Appearance: Multi-Scale Compensatory Codebooks for Talking Head Video Generation (CVPR 2025 · similarity 0.86)
- FaithfulFaces: Pose-Faithful Facial Identity Preservation for Text-to-Video Generation (NeurIPS 2026 · similarity 0.85)
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.
