Research map / Video / Video segmentation and tracking
AlphaVC: High-Performance and Efficient Learned Video Compression
Yibo Shi, Yunying Ge, Jing Wang, Jue Mao
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 cI-frame as first frame of each GoP after the first (ablation) | AlphaVC-P · HEVC Class B | BD-rate (PSNR): 21.4% -> 0% | Improved |
| add P-frames between cI-frames (GoP structure) (ablation) | AlphaVC-cI · HEVC Class B | BD-rate (PSNR): 92.7% -> 0% | Improved |
| add pixel-to-feature motion prediction (ablation) | AlphaVC-P · HEVC Class B | BD-rate (PSNR): 10.4% -> 0% | Improved |
| add Skip in motion latents (ablation) | AlphaVC-P without P2F MP · HEVC Class B | BD-rate (PSNR): 18.6% -> 10.4% | Improved |
| add Skip in residual latents (ablation) | AlphaVC-P without P2F MP and motion Skip · HEVC Class B | BD-rate (PSNR): 37.5% -> 18.6% | Improved |
| add probability-based entropy skipping (ablation) | AlphaVC · 1080p complexity | entropy coding runtime: Skip ratio of motion latents is as high as 90% | Improved |
| use AlphaVC instead of VTM-IPP | VTM-IPP · 5 test datasets | BD-rate (PSNR): -28.2% | Improved |
| use AlphaVC instead of VTM-LDP | VTM-LDP · 5 test datasets | BD-rate (PSNR): -23.5% -> -28.2% | Improved |
| use AlphaVC instead of TCMVC | TCMVC · 5 test datasets | BD-rate (MS-SSIM): -44.5% -> -52.2% | Improved |
| use AlphaVC instead of VTM for encoding | VTM-10.0-IPP · 1080p | encoding time: 661.9 -> 0.715 | Improved |
| use AlphaVC instead of TCMVC | TCMVC · 1080p | decoding time: 0.472 -> 0.379 | Improved |
| use AlphaVC instead of VTM for decoding | VTM-10.0-IPP · 1080p | decoding time: 0.224 -> 0.379 | Hurt |
Closest papers by meaning
- Neural Video Compression with Feature Modulation (CVPR 2024 · similarity 0.88)
- NVRC: Neural Video Representation Compression (NeurIPS 2024 · similarity 0.86)
- High Dynamic Range Video Compression: A Large-Scale Benchmark Dataset and A Learned Bit-depth Scalable Compression Algorithm (CVPR 2025 · similarity 0.86)
- Towards Practical Real-Time Neural Video Compression (CVPR 2025 · similarity 0.86)
- PNVC: Towards Practical INR-based Video Compression (AAAI 2025 · similarity 0.86)
- ECVC: Exploiting Non-Local Correlations in Multiple Frames for Contextual Video Compression (CVPR 2025 · similarity 0.86)
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.
