Research map / Image generation / Image restoration
The Last Byte: Learning Just Enough for Machine-Oriented Image Compression
Wuyuan Xie, Zhenming Li, Ye Liu, Jian Jin, Yun Song, Miaohui Wang
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 detector confidence map as input to the original image (ablation) | MVRNet · MVRSet test | E_total: 13.785 -> 7.735 | Improved |
| add original image to confidence-map-only input (ablation) | MVRNet · MVRSet test | E_total: 8.330 -> 7.735 (E_total); E_s 7.728 -> 8.250 | Mixed |
| add CAMM channel attention (ablation) | MVRNet · MVRSet test | E_total: 8.437 -> 7.844 | Improved |
| add CSFM fusion module (ablation) | MVRNet · MVRSet test | E_total: 8.437 -> 8.137 | Improved |
| combine CAMM and CSFM (ablation) | MVRNet · MVRSet test | E_total: 7.844 -> 7.735 | Improved |
| use single-pass no-reference MVRNet instead of OWJRD | OWJRD · MVRSet test | runtime: 7116.19s -> 0.17s | Improved |
| use MVRNet instead of JRDVCM | JRDVCM · MVRSet test | E_total: 4.655 -> 4.140 | Improved |
| use MVRNet vs H.266/VVC at COCO detection | H.266/VVC · COCO | mAP@.50 / BPP: 5.8720 BPP/0.7290 -> 0.5314 BPP/0.8640 mAP | Improved |
| use MVRNet with DETR detector vs H.266/VVC | H.266/VVC · COCO detection (DETR) | mAP@.50: 0.7290 -> 0.7100 mAP@.50, BPP 5.8720 -> 0.5435 | No clear effect |
| use MVRNet for instance segmentation vs H.266/VVC | H.266/VVC · COCO instance segmentation | mAP@.50: 0.6640 -> 0.6341 mAP@.50, BPP 5.8643 -> 0.4481 | Mixed |
Closest papers by meaning
- Firing Bits Where It Matters: Spiking-Guided Just Recognizable Distortion Modeling for Machine-Centric Video Coding (AAAI 2026 · similarity 0.87)
- Good, Cheap, and Fast: Overfitted Image Compression with Wasserstein Distortion (CVPR 2025 · similarity 0.85)
- Unified Coding for Both Human Perception and Generalized Machine Analytics with CLIP Supervision (AAAI 2025 · similarity 0.83)
- Test-Time Fine-Tuning of Image Compression Models for Multi-Task Adaptability (CVPR 2025 · similarity 0.83)
- Benchmarking and Enhancing VLM for Compressed Image Understanding (ICML 2026 · similarity 0.83)
- In defense of parameter sharing for model-compression (ICLR 2024 · similarity 0.83)
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.
