Research map / Architectures and efficiency / RNNs and normalization
Selective compression learning of latent representations for variable-rate image compression
Jooyoung Lee, Seyoon Jeong, Munchurl Kim
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 selective compression (vs w/o SC) (ablation) | SCR_Hyp (MSE) · Kodak | PSNR BD-rate: -4.30% | Improved |
| add selective compression (vs w/o SC) (ablation) | SCR_MS (MSE) · Kodak | PSNR BD-rate: -5.03% | Improved |
| add selective compression (vs w/o SC) (ablation) | SCR_Cxt (MSE) · Kodak | PSNR BD-rate: -1.18% | Improved |
| add selective compression (vs w/o SC) (ablation) | SCR_Hyp (MS-SSIM) · Kodak | MS-SSIM BD-rate: -4.03% | Improved |
| add selective compression (vs w/o SC) (ablation) | SCR_MS (MS-SSIM) · Kodak | MS-SSIM BD-rate: -3.00% | Improved |
| add selective compression (vs w/o SC) (ablation) | SCR_Cxt (MS-SSIM) · Kodak | MS-SSIM BD-rate: -1.22% | Improved |
| single SCR variable-rate model vs separately trained reference models | SCR_Hyp (MSE) · Kodak | PSNR BD-rate: -3.21% | Improved |
| single SCR variable-rate model vs separately trained reference models | SCR_MS (MSE) · Kodak | PSNR BD-rate: 1.30% | No clear effect |
| single SCR variable-rate model vs separately trained reference models | SCR_Cxt (MSE) · Kodak | PSNR BD-rate: 1.23% | No clear effect |
| add selective compression (vs w/o SC) (ablation) | SCR_Hyp · Kodak | decoding time: 11.28% | Improved |
| add selective compression (vs w/o SC) (ablation) | SCR_Cxt · Kodak | decoding time: 25.71% | Improved |
| add selective compression (vs w/o SC) (ablation) | SCR_MS · Kodak | encoding time: 6.96% | Improved |
Closest papers by meaning
- Variable Rate Deep Image Compression With a Conditional Autoencoder (ICCV 2019 · similarity 0.88)
- Conditional Probability Models for Deep Image Compression (CVPR 2018 · similarity 0.88)
- Improving Inference for Neural Image Compression (NeurIPS 2020 · similarity 0.87)
- Robustly overfitting latents for flexible neural image compression (NeurIPS 2024 · similarity 0.86)
- Conditional Latent Coding with Learnable Synthesized Reference for Deep Image Compression (AAAI 2025 · similarity 0.86)
- GLIC: General Format Learned Image Compression (AAAI 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.
