Research map / Video / Video segmentation and tracking
Neural Video Compression With Diverse Contexts
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 |
|---|---|---|---|
| Use DCVC-DC with diverse temporal and spatial contexts instead of DCVC-HEM | Neural video compression · RGB datasets average (BT.709) | BD-Rate (%) vs VTM-17.0, PSNR: 7.1 -> --17.8 | Helped |
| Use DCVC-DC instead of DCVC-HEM | Neural video compression · RGB datasets average (BT.709) | BD-Rate (%) vs VTM-17.0, MS-SSIM: --36.5 -> --47.6 | Helped |
| Use DCVC-DC instead of ECM-5.0 | Video compression · YUV420 datasets average | BD-Rate (%) vs VTM-17.0, PSNR: --13.1 -> --17.0 | Helped |
| Use DCVC-DC instead of DCVC-HEM | Neural video compression · RGB datasets average (BT.601) | BD-Rate (%) vs VTM-17.0, PSNR: 1.7 -> --18.0 | Helped |
| Add hierarchical quality structure during training (ablation) | DCVC-DC · HEVC datasets (RGB), ablation | BD-Rate (%): 8.4 -> 0.0 | Helped |
| Add cross-group interaction to offset diversity (ablation) | DCVC-DC · HEVC datasets (RGB), ablation | BD-Rate (%): 12.1 -> 8.4 | Helped |
| Add offset diversity without cross-group interaction (ablation) | DCVC-DC · HEVC datasets (RGB), ablation | BD-Rate (%): 14.7 -> 12.1 | Helped |
| Use quadtree partition-based entropy coding instead of the dual spatial model (ablation) | DCVC-DC · HEVC datasets (RGB), ablation | BD-Rate (%): 21.3 -> 14.7 | Helped |
| Use depthwise separable convolution instead of normal convolution (ablation) | DCVC-DC · HEVC datasets (RGB), structure ablation | BD-Rate (%): 3.5 -> 2.4 | Helped |
| Use unequal channel numbers for features at different resolutions (ablation) | DCVC-DC · HEVC datasets (RGB), structure ablation | BD-Rate (%): 2.4 -> 1.1 | Helped |
| Move partial quantization operations to higher resolution (ablation) | DCVC-DC · HEVC datasets (RGB), structure ablation | BD-Rate (%): 1.1 -> 0.0 | Helped |
| Use DCVC-DC instead of DCVC-HEM | Neural video compression · DCVC-HEM versus DCVC-DC complexity | MACs: 3279G -> 2642G | Helped |
Closest papers by meaning
- Neural Video Compression with Context Modulation (CVPR 2025 · similarity 0.91)
- Towards Practical Real-Time Neural Video Compression (CVPR 2025 · similarity 0.89)
- Neural Video Compression with Feature Modulation (CVPR 2024 · similarity 0.88)
- Augmented Deep Contexts for Spatially Embedded Video Coding (CVPR 2025 · similarity 0.88)
- NVRC: Neural Video Representation Compression (NeurIPS 2024 · similarity 0.87)
- PNVC: Towards Practical INR-based Video Compression (AAAI 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.