Research map / Architectures and efficiency / RNNs and normalization
HCF: Hierarchical Cascade Framework for Distributed Multi-Stage Image Compression
Junhao Cai, Taegun An, Chengjun Jin, Sung Il Choi, Juhyun Park, Changhee Joo
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 |
|---|---|---|---|
| quantize early (edge policy) instead of late (ablation) | HCF with MLIC++ · Kodak (MLIC++, nq=2, d=2) | PSNR (dB): 29.768 -> 30.264 at 0.1674 bpp | Improved |
| quantize early (edge policy) instead of late (ablation) | HCF with MLIC++ · Kodak (MLIC++, nq=4, d=1) | PSNR (dB): 27.929 -> 28.524 | Improved |
| use HCF latent-space transformation instead of pixel-domain recompression (DRF) | MLIC++ · Kodak | BD-Rate PSNR (%): DRF +4.87% worse than HCF | Improved |
| use HCF instead of progressive compression (Presta) | MLIC++ · Kodak | BD-Rate PSNR (%): Presta +12.64% worse than HCF | Improved |
| replace DRF with HCF for transform operations | cheng2020_attn · Node1 efficiency | FLOPs reduction: 97.8% FLOPs, 95.5% GPU memory, 90.0% time reduction | Improved |
| shorter compression path with HCF (5->1 vs 6->1) | HCF cheng2020_attn · CLIC2020-mobile | BD-Rate PSNR (%): -7.13 | Improved |
| remove either phi_inter or phi_intra transform module (ablation) | HCF · CLIC2020-mobile | rate-distortion | Hurt |
Closest papers by meaning
- DLF: Extreme Image Compression with Dual-generative Latent Fusion (ICCV 2025 · similarity 0.86)
- Differentiable Vector Quantization for Rate-Distortion Optimization of Generative Image Compression (CVPR 2026 · similarity 0.85)
- Split Hierarchical Variational Compression (CVPR 2022 · similarity 0.84)
- Ultra Lowrate Image Compression with Semantic Residual Coding and Compression-aware Diffusion (ICML 2025 · similarity 0.84)
- HyperVQ: Enabling Hyperprior Entropy Modeling for VQ-Based Generative Image Compression (NeurIPS 2026 · similarity 0.83)
- 4DGCPro: Efficient Hierarchical 4D Gaussian Compression for Progressive Volumetric Video Streaming (NeurIPS 2025 · 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.
