Research map / Video / Video generation
Ariadne's Thread of LipSync: Unraveling Forgeries via Inconsistency between Lip Motions and Head Poses
Tianyi She, Jiawei Liu, Weifeng Liu, Hanqing Zhao, Weiming Zhang, Kejiang Chen
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 alignment loss between lip and pose embeddings (ablation) | LipDA Stage I · AVLips | ACC: 80.56 -> 97.90 | Improved |
| Add head pose branch (ablation) | LipDA Stage I · AVLips | ACC: 74.78 -> 97.90 | Improved |
| Add lip branch (ablation) | LipDA Stage I · AVLips | ACC: 53.50 -> 97.90 | Improved |
| Replace ResNet-18 lip encoder with CLIP-ViT (ablation) | LipDA · AVLips | ACC: 97.90 -> 85.73 | Hurt |
| Replace ResNet-18 lip encoder with MobileNet (ablation) | LipDA · AVLips | ACC: 97.90 -> 96.23 | Hurt |
| Use flatten or pooling pose features instead of LSTM (ablation) | LipDA · AVLips | ACC: 97.90 -> 95.27 (Flatten); 97.90 -> 97.20 (Pooling) | Hurt |
| Add audio MSM score to visual-only LipDA (ablation) | LipDA · TalkHeadBench | AUC: 98.65 -> 97.50 | Mixed |
| Add audio MSM score to visual-only LipDA (ablation) | LipDA · LipSync-A | AUC: 97.83 -> 99.42 | Improved |
| Increase sliding window length T excessively (ablation) | LipDA · detection accuracy vs window T | accuracy | Hurt |
| Increase total video length up to about 8 seconds (ablation) | LipDA · detection accuracy vs video length | accuracy | Improved |
| Use landmark-based pose features under Gaussian noise perturbation (ablation) | LipDA · AVLips Level-5 Gaussian noise | AUC: 99.72 -> 90.98 | Hurt |
| Use LipDA under Level-3 visual corruptions | LipDA · AVLips | AUC (average): 85.53 -> 99.11 (vs LipForensics) | Improved |
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.
- SpeechForensics — SpeechForensics: Audio-Visual Speech Representation Learning for Face Forgery Detection (NeurIPS 2024) · better on 7 of 8 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
- Lips Are Lying: Spotting the Temporal Inconsistency between Audio and Visual in Lip-Syncing DeepFakes (NeurIPS 2024 · similarity 0.92)
- OmniSync: Towards Universal Lip Synchronization via Diffusion Transformers (NeurIPS 2025 · similarity 0.91)
- StyleSync: High-Fidelity Generalized and Personalized Lip Sync in Style-Based Generator (CVPR 2023 · similarity 0.90)
- Perceptually Accurate 3D Talking Head Generation: New Definitions, Speech-Mesh Representation, and Evaluation Metrics (CVPR 2025 · similarity 0.87)
- DiVAS: Video and Audio Synchronization with Dynamic Frame Rates (CVPR 2024 · similarity 0.87)
- PointTalk: Audio-Driven Dynamic Lip Point Cloud for 3D Gaussian-based Talking Head Synthesis (AAAI 2025 · 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.
