Research map / Video / Video segmentation and tracking
HyTIP: Hybrid Temporal Information Propagation for Masked Conditional Residual Video Coding
Yi-Hsin Chen, Yi-Chen Yao, Kuan-Wei Ho, Chun-Hung Wu, Huu-Tai Phung, Martin Benjak, Jörn Ostermann, Wen-Hsiao Peng
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 hybrid buffering (explicit flow + small implicit features) to motion coding instead of explicit only (ablation) | HyTIP · Average, PSNR-RGB BD-rate | BD-rate (%): 0 -> -14.7 | Improved |
| use hybrid instead of implicit buffering for motion at small buffer (ablation) | HyTIP motion coding · Average, PSNR-RGB BD-rate | BD-rate (%): -1.3 -> -14.7 | Improved |
| reduce implicit motion buffer to a small size (ablation) | Implicit motion · Average, PSNR-RGB BD-rate | BD-rate (%): -12.2 -> -1.3 | Hurt |
| reduce hybrid motion buffer from 4 to 0.125 feature maps (ablation) | Hybrid motion · Average, PSNR-RGB BD-rate | BD-rate (%): -16.0 -> -14.7 | No clear effect |
| use hybrid instead of explicit inter-frame buffer (ablation) | HyTIP inter-frame · Average, PSNR-RGB BD-rate | BD-rate (%): -14.7 -> -21.5 | Improved |
| shrink implicit inter-frame buffer from 51 to 4 maps (ablation) | Implicit inter-frame · Average, PSNR-RGB BD-rate | BD-rate (%): -19.6 -> -10.2 | Hurt |
| shrink hybrid inter-frame buffer from 48 to 2 maps (ablation) | Hybrid inter-frame · Average, PSNR-RGB BD-rate | BD-rate (%): -21.9 -> -21.5 | No clear effect |
| train with 10-frame sequences instead of 5 for explicit buffering (ablation) | Explicit inter-frame · BVI-DVC, BD-rate | BD-rate (%): -13.2 -> -16.1 | Improved |
| train with 10-frame sequences for implicit buffering (ablation) | Implicit inter-frame · BVI-DVC, BD-rate | BD-rate (%): -13.6 -> -18.5 | Improved |
| train with 10-frame sequences for hybrid buffering (ablation) | HyTIP · BVI-DVC, BD-rate | BD-rate (%): -20.0 -> -24.7 | Improved |
| HyTIP vs DCVC-DC on HEVC-E conferencing sequences | HyTIP · HEVC-E PSNR-RGB | BD-rate (%): -28.1 -> -21.2 | Hurt |
| HyTIP vs DCVC-DC on average PSNR | HyTIP · PSNR-RGB average | BD-rate (%): -23.6 -> -26.5 | Improved |
Closest papers by meaning
- Learned Video Compression via Joint Spatial-Temporal Correlation Exploration (AAAI 2020 · similarity 0.88)
- MH-LVC: Multi-Hypothesis Temporal Prediction for Learned Conditional Residual Video Coding (ICCV 2025 · similarity 0.88)
- Motion Information Propagation for Neural Video Compression (CVPR 2023 · similarity 0.88)
- Neural Video Compression with Context Modulation (CVPR 2025 · similarity 0.87)
- Neural Inter-Frame Compression for Video Coding (ICCV 2019 · similarity 0.87)
- Neural Video Compression with Feature Modulation (CVPR 2024 · 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.
