Research map / Video / Video segmentation and tracking
Perceptual Neural Video Compression with Color Separation and Rank Chain
Xiongzhuang Liang, Chuanbo Tang, Zhuoyuan Li, Li Li, Dong Liu
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 color separation (separate Luma-Net and Chroma-Net) to DCVC-DC (ablation) | DCVC-DC (re-implemented) · BD-rate ablation | YUV PSNR BD-rate (%): 0 -> -7.2 | Improved |
| add Color Fidelity Refinement (CFR) module (ablation) | PNVC-C · BD-rate ablation | YUV PSNR BD-rate (%): -7.2 -> -13.2 | Improved |
| extend training sequence length (long-sequence training) (ablation) | PNVC-C · BD-rate ablation | YUV PSNR BD-rate (%): -13.2 -> -21.4 | Improved |
| replace ResBlocks with Partial Conv ResBlocks (ablation) | PNVC-C · BD-rate ablation | YUV PSNR BD-rate (%): -21.4 -> -20.7; 1583.6 -> 1412.6 kMacs/pixel | Mixed |
| add LPIPS loss alone to PNVC-C-Base (ablation) | PNVC-C-Base · BD-rate ablation | LPIPS BD-rate (%): LPIPS 0 -> -51.2; YUV PSNR 0 -> 14.7 | Mixed |
| add standard WGAN loss alongside LPIPS (ablation) | PNVC-C-Base · BD-rate ablation | LPIPS BD-rate (%): -51.2 -> -60.1; YUV PSNR 14.7 -> 11.5 | Improved |
| use Rc-GAN rank chain loss instead of standard WGAN (ablation) | PNVC-C-Base · BD-rate ablation | LPIPS BD-rate (%): -60.1 -> -77.5; YUV PSNR 11.5 -> 8.3 | Improved |
| apply Rc-GAN perceptual fine-tuning to PNVC-C-Base | PNVC-C-Base · Average of 7 datasets | YUV PSNR BD-rate (%): -25.26 -> -15.91 | Hurt |
| PNVC-CR versus DiffVC on HEVC B | PNVC-CR · HEVC B, IP=32 | LPIPS: 0.0707 (0.1911 bpp) -> 0.0455 (0.0523 bpp) | Improved |
Closest papers by meaning
- Neural Video Compression with Feature Modulation (CVPR 2024 · similarity 0.89)
- PNVC: Towards Practical INR-based Video Compression (AAAI 2025 · similarity 0.87)
- Towards Practical Real-Time Neural Video Compression (CVPR 2025 · similarity 0.87)
- NVRC: Neural Video Representation Compression (NeurIPS 2024 · similarity 0.86)
- C3: High-Performance and Low-Complexity Neural Compression from a Single Image or Video (CVPR 2024 · similarity 0.86)
- Combining Frame and GOP Embeddings for Neural Video Representation (CVPR 2024 · 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.
