Research map / Video / Video segmentation and tracking
MH-LVC: Multi-Hypothesis Temporal Prediction for Learned Conditional Residual Video Coding
Huu-Tai Phung, Zong-Lin Gao, Yi-Chen Yao, Kuan-Wei Ho, Yi-Hsin Chen, Yu-Hsiang Lin, Alessandro Gnutti, 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 spatial gating to long/short-term feature fusion (ablation) | MH-LVC · HEVC Class B-E | BD-rate PSNR-RGB (%): -4.2, -5.9, -5.5, -13.1 vs 0.0 (Class B, C, D, E) | Improved |
| use spatial gating instead of CNN-based fusion (ablation) | MH-LVC LSTP · HEVC Class B | BD-rate PSNR-RGB (%): -2.1 -> -4.2 | Improved |
| use spatial gating instead of CNN-based fusion (ablation) | MH-LVC LSTP · HEVC Class E | BD-rate PSNR-RGB (%): -9.1 -> -13.1 | Improved |
| add a second long-term key frame (MH-LVC-2 vs MH-LVC-1) (ablation) | MH-LVC · UVG | BD-rate PSNR-RGB (%): -9.6 -> -12.5 | Improved |
| add a second long-term key frame (MH-LVC-2 vs MH-LVC-1) (ablation) | MH-LVC · HEVC-B | BD-rate PSNR-RGB (%): -12.8 -> -15.7 | Improved |
| add a third long-term key frame beyond two (ablation) | MH-LVC · HEVC-B | BD-rate PSNR-RGB (%): -3.4 -> -3.8 | No clear effect |
| add a third long-term key frame beyond two (encoding cost) (ablation) | MH-LVC · HEVC-B | Encoding kMAC/pixel: 2611 -> 3716 kMAC/pixel | Hurt |
| increase mini-GOP size from 4 to 16 (ablation) | MH-LVC · HEVC-B | BD-rate PSNR-RGB (%): 0.0 -> 2.9 | Hurt |
| increase mini-GOP size from 4 to 16 (ablation) | MH-LVC · UVG | BD-rate PSNR-RGB (%): 0.0 -> 2.1 | Hurt |
| use long-short prediction (LS) instead of short-short (SS) (ablation) | MH-LVC · HEVC-B | BD-rate PSNR-RGB (%): 14.6 -> 0.0 | Improved |
| use long-short prediction (LS) instead of two previous frames (TP) (ablation) | MH-LVC · HEVC-B | BD-rate PSNR-RGB (%): 3.3 -> 0.0 | Improved |
| use explicit two-frame buffering (MH-LVC-1) instead of implicit TCM buffering (ablation) | MH-LVC-1 · UVG, HEVC-B, MCL-JCV | BD-rate PSNR-RGB (%): 7.8, 3.2, 8.9 BD-rate for TCM vs 0.0 | Improved |
Closest papers by meaning
- M-LVC: Multiple Frames Prediction for Learned Video Compression (CVPR 2020 · similarity 0.90)
- ECVC: Exploiting Non-Local Correlations in Multiple Frames for Contextual Video Compression (CVPR 2025 · similarity 0.88)
- Neural Video Compression with Context Modulation (CVPR 2025 · similarity 0.86)
- Neural Video Compression with Feature Modulation (CVPR 2024 · similarity 0.85)
- AdaCodec: A Predictive Visual Code for Video MLLMs (NeurIPS 2026 · similarity 0.85)
- Augmented Deep Contexts for Spatially Embedded Video Coding (CVPR 2025 · similarity 0.84)
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.
