Research map / Architectures and efficiency / Quantization
Learning Scalable lY=-Constrained Near-Lossless Image Compression via Joint Lossy Image and Residual Compression
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Jointly train lossy-image and residual compression | Near-lossless codec · DIV2K | bpsp: 2.91 -> 2.81 | Helped |
| Use the proposed codec instead of FLIF | Near-lossless codec · CLIC.p | bpsp: 2.72 -> 2.66 | Helped |
| Use the proposed codec instead of FLIF | Near-lossless codec · Kodak | bpsp: 2.90 -> 3.04 | Hurt |
| Quantize residuals for τ=1 rather than lossless coding at τ=0 | Ours · Kodak | bit rate: about 40% | Helped |
| Add bias correction to quantized-residual probability inference (ablation) | Ours · Kodak | bpsp | Helped |
| Set λ from 0 to 0.03 (ablation) | Ours · Kodak | Total Rate: 2.93 -> 3.04 | Hurt |
| Set λ from 0 to 0.03 (ablation) | Ours · Kodak | lossy reconstruction PSNR: 24.09 -> 39.29 | Helped |
| Set λ to 0.05 rather than 0.03 (ablation) | Ours · Kodak | Total Rate: 3.04 -> 3.06 | Hurt |
| Set λ to 0.05 rather than 0.03 (ablation) | Ours · Kodak | lossy reconstruction PSNR: 39.29 -> 39.91 | Helped |
| Set λ to 0 rather than 0.03 (ablation) | Ours for τ>0 · Kodak | near-lossless reconstruction visual quality | Hurt |
| Use the proposed codec instead of the best listed baseline at τ=1 | Ours w/ bc · Kodak | bpsp: 2.41 -> 1.84 | Helped |
| Use the proposed codec instead of the best listed baseline at τ=2 | Ours w/ bc · Kodak | bpsp: 2.01 -> 1.38 | Helped |
Closest papers by meaning
- Learning Better Lossless Compression Using Lossy Compression (CVPR 2020 · similarity 0.90)
- Practical Full Resolution Learned Lossless Image Compression (CVPR 2019 · similarity 0.90)
- Improving Inference for Neural Image Compression (NeurIPS 2020 · similarity 0.88)
- Efficient Learned Image Compression without Entropy Coding (ICML 2026 · similarity 0.88)
- Large Language Models for Lossless Image Compression: Next-Pixel Prediction in Language Space is All You Need (NeurIPS 2025 · similarity 0.86)
- Ultra Lowrate Image Compression with Semantic Residual Coding and Compression-aware Diffusion (ICML 2025 · 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.