Research map / Image generation / Image restoration
Cross-Granularity Online Optimization with Masked Compensated Information for Learned Image Compression
Haowei Kuang, Wenhan Yang, Zongming Guo, Jiaying Liu
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 DDT module with adaptive QP selection to the baseline (ablation) | CGOO · Kodak | R-D performance (PSNR vs bpp) | Improved |
| add MSC module with masked selection on top of DDT (ablation) | CGOO · Kodak | R-D performance (PSNR vs bpp) | Improved |
| add adaptive QP selection to the DDT module (ablation) | CGOO · Kodak | R-D performance (PSNR vs bpp) | Improved |
| add mask-based selection to the MSC module (ablation) | CGOO · Kodak | R-D performance (PSNR vs bpp) | Improved |
| add cross-granularity online optimization to Liu (CVPR-23) baseline | CGOO · CLIC Professional validation | BD-rate vs BPG: -41.65% -> -43.86% | Improved |
| use CGOO instead of Li (ICLR-24) larger model | CGOO · CLIC Professional validation | BD-rate vs BPG: -42.80% -> -43.86% | Improved |
| add online optimization to baseline at encoding time | CGOO · Time complexity, 768x512 image | encoding time: 0.25s -> 36.34s | Hurt |
| add online optimization to baseline at decoding time | CGOO · Time complexity, 768x512 image | decoding time: 0.24s -> 0.25s | No clear effect |
Closest papers by meaning
- Once-for-All: Controllable Generative Image Compression with Dynamic Granularity Adaptation (ICLR 2025 · similarity 0.88)
- Learning End-to-End Lossy Image Compression: A Benchmark (TPAMI 2021 · similarity 0.88)
- Balanced Rate-Distortion Optimization in Learned Image Compression (CVPR 2025 · similarity 0.86)
- Outlier-aware Slicing for Post-Training Quantization in Vision Transformer (ICML 2024 · similarity 0.85)
- Compression of Structured Data with Autoencoders: Provable Benefit of Nonlinearities and Depth (ICML 2024 · similarity 0.85)
- Learned Image Compression with Dictionary-based Entropy Model (CVPR 2025 · similarity 0.85)
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.
