Research map / Video / Video segmentation and tracking
Video Compression With Entropy-Constrained Neural Representations
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Replace NeRV's fully connected spatial projection with a fully convolutional architecture (AdaIN blocks, single FC for temporal input) | NeRV-style frame-based INR · BigBuckBunny representational capacity (12.5M params) | PSNR: 41.1 -> 42.7 | No clear effect |
| Fully convolutional architecture vs NeRV decoding speed | NeRV-style frame-based INR · BigBuckBunny representational capacity | FPS: 88 (NeRV) -> 56 (ours) -> 45.3 (E-NeRV) | Mixed |
| Entropy minimization (joint rate-distortion training with quantization-aware training and entropy coding) added to NeRV | NeRV · UVG | R-D performance (PSNR, MS-SSIM vs bpp) | Helped |
| Joint rate-distortion entropy minimization in place of post-training quantization/pruning | proposed INR video compression · UVG | R-D performance vs HEVC and DVC | Helped |
| Entropy minimization applied as fine-tuning of a pretrained INR (0.8e epochs with lambda=0, then target lambda) | proposed INR video compression · UVG | R-D performance | No clear effect |
| Per-layer independent entropy modeling of weights vs joint modeling of all weights | proposed INR video compression · UVG | R-D performance | Helped |
| Using a larger architecture (50M) with a smaller (14M) architecture at different bpp ranges, vs one architecture for all rates | proposed INR video compression · UVG | R-D performance | Helped |
Closest papers by meaning
- NVRC: Neural Video Representation Compression (NeurIPS 2024 · similarity 0.93)
- Combining Frame and GOP Embeddings for Neural Video Representation (CVPR 2024 · similarity 0.90)
- Neural Video Compression with Feature Modulation (CVPR 2024 · similarity 0.90)
- Towards Practical Real-Time Neural Video Compression (CVPR 2025 · similarity 0.89)
- On Quantizing Neural Representation for Variable-Rate Video Coding (ICLR 2025 · similarity 0.87)
- PNVC: Towards Practical INR-based Video Compression (AAAI 2025 · similarity 0.87)
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.