Research map / Image generation / Image restoration
Content-Aware Mamba for Learned Image Compression
Yunuo Chen, Zezheng Lyu, Bing He, Hongwei Hu, Qi Wang, Yuan Tian, Li Song, Wenjun Zhang, Guo Lu
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 content-adaptive token permutation (CTP) to vanilla single-scan Mamba (ablation) | CMIC · Kodak | BD-rate (%): -13.26 -> -15.21 | Improved |
| add global-prior prompting (GPP) to vanilla single-scan Mamba (ablation) | CMIC · Kodak | BD-rate (%): -13.26 -> -14.27 | Improved |
| combine CTP and GPP (ablation) | CMIC · Tecnick | BD-rate (%): -17.74 -> -21.34 | Improved |
| replace 2D multi-directional Mamba with CAM blocks (ablation) | CMIC · Kodak | BD-rate (%): -14.13 -> -15.91 | Improved |
| use CAM blocks together with window attention instead of CAM-only or attention-only (ablation) | CMIC · Kodak | BD-rate (%): -14.68 -> -15.91 | Improved |
| increase cluster count K from 64 to 128 (ablation) | CMIC · Kodak | BD-rate (%): -15.91 -> -15.96 | No clear effect |
| decrease cluster count K from 64 to 32 (ablation) | CMIC · Kodak | BD-rate (%): -15.91 -> -14.97 | Hurt |
| use CAM block in the entropy model (ablation) | CMIC · Kodak | BD-rate (%): -15.91 -> -15.81 | No clear effect |
| add CTP and GPP modules (ablation) | CMIC · 2K decoding | decoding latency (s): 0.387s -> 0.405s | Hurt |
| use CMIC transforms with FTIC's T-CA entropy model (ablation) | CMIC · Kodak | BD-rate (%): -12.94 -> -15.67 | Improved |
Closest papers by meaning
- MambaIC: State Space Models for High-Performance Learned Image Compression (CVPR 2025 · similarity 0.90)
- CALLIC: Content Adaptive Learning for Lossless Image Compression (AAAI 2025 · similarity 0.86)
- SaMam: Style-aware State Space Model for Arbitrary Image Style Transfer (CVPR 2025 · similarity 0.85)
- Frequency-Aware Transformer for Learned Image Compression (ICLR 2024 · similarity 0.85)
- MobileMamba: Lightweight Multi-Receptive Visual Mamba Network (CVPR 2025 · similarity 0.85)
- Linear Attention Modeling for Learned Image Compression (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.
