Research map / Video / Video segmentation and tracking
A Coding Framework and Benchmark Towards Low-Bitrate Video Understanding
Yuan Tian, Guo Sheng Lu, Yichao Yan, Guangtao Zhai, Li Chen, Zhiyong Gao
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 VCS semantic stream on top of VVC | VVC codec · MOT17 (@0.01bpp) | MOTA: 64.86 -> 67.79 | Improved |
| add VCS semantic stream on top of VVC | VVC codec · DAVIS2017 VOS (@0.01bpp) | J&F: 67.47 -> 72.44 | Improved |
| add VCS to VVC with TSM on action recognition | VVC codec · UCF101 action recognition (TSM, 0.04bpp) | Top-1 accuracy (%): 76.97 -> 88.19 | Improved |
| apply the VVC-trained VCS to H.264 without adaptation | H.264 · UCF101 action recognition (TSM) | Top-1 accuracy (%): +13.61 Top1, -37.92 Bit% | Improved |
| apply VCS to H.264 on MOT | H.264 · MOT17 | MOTA: +3.05% MOTA, -22.75 bit | Improved |
| apply VCS to H.264 on VOS | H.264 · DAVIS2017 VOS | J&F: +6.02 J&F, -36.51 bit | Improved |
| add label fine-tuning of VCS with downstream labels | VCS · UCF101 (SlowFast, 0.02bpp) | Top-1 accuracy (%): 85.04 -> 87.14 | Improved |
| add label fine-tuning of VCS with downstream labels | VCS · HMDB51 (SlowFast, 0.02bpp) | Top-1 accuracy (%): 63.62 -> 64.95 | Improved |
| add the semantic stream (vs removing the semantic encoder, LFN-Only) (ablation) | VCS* · HEVC Class C (@0.10bpp) | PSNR (dB): 32.64 -> 32.89 | Improved |
| add semantic stream instead of BasicVSR++ enhancement (ablation) | VCS* · HEVC Class C (@0.10bpp) | PSNR (dB): 32.68 -> 32.89 | Improved |
| add PSNR-oriented decoder tuning (TuneDec) to vanilla VCS (ablation) | VCS · HEVC Class C (@0.10bpp) | PSNR (dB): 32.38 -> 32.83 | Improved |
| add encoder-and-decoder rate-distortion fine-tuning (VCS*) over decoder-only tuning (ablation) | VCS-TuneDec · HEVC Class C (@0.10bpp) | PSNR (dB): 32.83 -> 32.89 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- BasicVSR++ — BasicVSR: The Search for Essential Components in Video Super-Resolution and Beyond (CVPR 2021) · better on 1 of 1 shared result
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
- Non-Semantics Suppressed Mask Learning for Unsupervised Video Semantic Compression (ICCV 2023 · similarity 0.90)
- One Token per Highly Selective Frame: Towards Extreme Compression for Long Video Understanding (NeurIPS 2025 · similarity 0.87)
- Neural Video Compression with Context Modulation (CVPR 2025 · similarity 0.87)
- Towards Practical Real-Time Neural Video Compression (CVPR 2025 · similarity 0.86)
- Video-XL: Extra-Long Vision Language Model for Hour-Scale Video Understanding (CVPR 2025 · similarity 0.86)
- VideoPrism: A Foundational Visual Encoder for Video Understanding (ICML 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.
