Research map / Video / Video segmentation and tracking
Enhanced Neural Video Representation Compression with High Scalability
Ho Man Kwan, Tianhao Peng, Fan Zhang, Mike Nilsson, Andrew Gower, David Bull
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 multiple high-resolution feature grids (ablation) | NVRC++ (S2) · UVG (anchor NVRC++ S2) | BD-rate PSNR increase when removed: 44.01 | Improved |
| add random feature grid masking (ablation) | NVRC++ (S2) · UVG (anchor NVRC++ S2) | BD-rate PSNR increase when removed: 30.33 | Improved |
| add spatial prior to entropy model (ablation) | NVRC++ (S2) · UVG (anchor NVRC++ S2) | BD-rate PSNR increase when removed: 1.42 | Improved |
| add temporal prior to entropy model (ablation) | NVRC++ (S2) · UVG (anchor NVRC++ S2) | BD-rate PSNR increase when removed: 5.13 | Improved |
| add scale prior to entropy model (ablation) | NVRC++ (S2) · UVG (anchor NVRC++ S2) | BD-rate PSNR increase when removed: 6.12 | Improved |
| use multi-reference input instead of 2D stem (ablation) | NVRC++ (S2) · UVG (anchor NVRC++ S2) | BD-rate PSNR increase when 2D stem used: 35.55 | Improved |
| use multi-reference input instead of 3D stem (ablation) | NVRC++ (S2) · UVG (anchor NVRC++ S2) | BD-rate PSNR increase when 3D stem used: 7.76 | Improved |
| use HiNeRV++ (high-resolution grids, light layers) instead of HiNeRV | HiNeRV · UVG | PSNR (dB) at 300 epochs: 35.27 -> 39.46 | Improved |
| use HiNeRV++ instead of HiNeRV | HiNeRV · UVG | decoding FPS: 35.5 -> 162.9 | Improved |
| use NVRC++ S3 instead of NVRC | NVRC · MCL-JCV | BD-rate PSNR vs x265 (%): -51.61 -> -55.90 | Improved |
| scale NVRC++ from S3 to S4 | NVRC++ · MCL-JCV | BD-rate PSNR vs x265 (%): -55.90 -> -51.84 | Hurt |
| scale NVRC++ from S3 to S4 | NVRC++ · UVG | BD-rate PSNR vs x265 (%): -71.20 -> -74.82 | 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.
- NVRC — NVRC: Neural Video Representation Compression (NeurIPS 2024) · better on 3 of 3 shared results
- HiNeRV — HiNeRV: Video Compression with Hierarchical Encoding-based Neural Representation (NeurIPS 2023) · better on 3 of 3 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
- NVRC: Neural Video Representation Compression (NeurIPS 2024 · similarity 0.95)
- PNVC: Towards Practical INR-based Video Compression (AAAI 2025 · similarity 0.91)
- Towards Practical Real-Time Neural Video Compression (CVPR 2025 · similarity 0.91)
- Boosting Neural Representations for Videos with a Conditional Decoder (CVPR 2024 · similarity 0.90)
- Neural Video Compression with Feature Modulation (CVPR 2024 · similarity 0.90)
- SINR: Sparsity Driven Compressed Implicit Neural Representations (CVPR 2025 · similarity 0.89)
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.
