Research map / Video / Video segmentation and tracking
Audio-Assisted Face Video Restoration with Temporal and Identity Complementary Learning
Yuqin Cao, Yixuan Gao, Wei Sun, Xiaohong Liu, Yulun Zhang, Xiongkuo Min
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 identity features (landmark + audio intra-frame module) (ablation) | GAVN · VoxCeleb2 compression artifact removal | PSNR: 28.7797 -> 28.9780 | Improved |
| add identity features (ablation) | GAVN · VoxCeleb2 deblur | PSNR: 38.8352 -> 39.3441 | Improved |
| add identity features (ablation) | GAVN · VoxCeleb2 super-resolution | PSNR: 35.7568 -> 36.1462 | Improved |
| add audio features (ablation) | GAVN · VoxCeleb2 compression artifact removal | PSNR: 28.9480 -> 28.9780 | Improved |
| add audio features (ablation) | GAVN · VoxCeleb2 deblur | PSNR: 39.0181 -> 39.3441 | Improved |
| add audio features (ablation) | GAVN · VoxCeleb2 super-resolution | PSNR: 35.8427 -> 36.1462 | Improved |
| add audio features (ablation) | GAVN · VoxCeleb2 compression artifact removal | LPIPS: 0.1718 -> 0.1658 | Improved |
| GAVN vs audio-aided DAVD-Net | GAVN · VoxCeleb2 compression artifact removal | PSNR: 28.7269 -> 28.9780 | Improved |
| GAVN vs audio-aided DAVD-Net | GAVN · Obama compression artifact removal | PSNR: 28.6645 -> 30.0221 | Improved |
| GAVN vs BasicVSR++ at low resolution scale 2 | GAVN · VoxCeleb2 super-resolution distortion level | PSNR: 42.4803 -> 44.1499 | Improved |
| stronger compression (CRF 35 vs 45) gain of GAVN over DAVD-Net | GAVN · VoxCeleb2 compression | PSNR: +0.1130 at CRF 35; +0.2923 at CRF 45 | Mixed |
| GAVN on real-world degraded videos vs DAVD-Net | GAVN · Real-world videos | SyncNet confidence: 1.6581 -> 1.7895 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- DAVD-Net — DAVD-Net: Deep Audio-Aided Video Decompression of Talking Heads (CVPR 2020) · better on 11 of 11 shared results
- BasicVSR++ — BasicVSR: The Search for Essential Components in Video Super-Resolution and Beyond (CVPR 2021) · better on 6 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
- Multi-Modality Deep Restoration of Extremely Compressed Face Videos (TPAMI 2022 · similarity 0.88)
- Efficient Video Face Enhancement with Enhanced Spatial-Temporal Consistency (CVPR 2025 · similarity 0.87)
- SVFR: A Unified Framework for Generalized Video Face Restoration (CVPR 2025 · similarity 0.85)
- FaceMe: Robust Blind Face Restoration with Personal Identification (AAAI 2025 · similarity 0.84)
- Dynamic Content Prediction with Motion-aware Priors for Blind Face Video Restoration (CVPR 2025 · similarity 0.84)
- Discrete Prior-Based Temporal-Coherent Content Prediction for Blind Face Video Restoration (AAAI 2025 · similarity 0.83)
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.
