Research map / 3D vision / 3D generation
Parallax to Align Them All: An OmniParallax Attention Mechanism for Distributed Multi-View Image Compression
Haotian Zhang, Feiyue Long, Yixin Yu, Jian Xue, Haocheng Tang, Tongda Xu, Zhenning Shi, Yan Wang, Siwei Ma, Jiaqi Zhang
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 |
|---|---|---|---|
| use OPAM instead of JCT fusion (ablation) | ParaHydra · WildTrack(3) | bitrate: +5.64% bitrate without OPAM | Improved |
| use OPAM instead of 2D self-attention (ablation) | ParaHydra · WildTrack(3) | bitrate: +11.30% bitrate with 2D Attn | Improved |
| add horizontal parallax attention to OPAM (ablation) | ParaHydra · WildTrack(3) | bitrate: +8.79% bitrate without HPA | Improved |
| add vertical parallax attention to OPAM (ablation) | ParaHydra · WildTrack(3) | bitrate: +3.63% bitrate without VPA | Improved |
| use parallax channel context module (PCCM) (ablation) | ParaHydra · WildTrack(3) | bitrate: +4.32% bitrate without PCCM | Improved |
| use parallax global context module (PGCM) (ablation) | ParaHydra · WildTrack(3) | bitrate: +2.92% bitrate without PGCM | Improved |
| use Para-EM instead of MLIC entropy model (ablation) | ParaHydra · WildTrack(3) | bitrate: +5.13% bitrate with MLIC | Improved |
| give encoder access to other views (joint encoding) (ablation) | ParaHydra · WildTrack(3) | PSNR: +0.01 dB PSNR | No clear effect |
| remove inter-view access at decoder (separate encoding and decoding) (ablation) | ParaHydra · WildTrack(3) | PSNR: -0.54 dB PSNR | Hurt |
| increase number of input views from 3 to 6 | ParaHydra · WildTrack | BDBR: -19.72% -> -24.18% BDBR | Improved |
| increase number of input views from 3 to 4 vs LMVIC | ParaHydra · Mip-NeRF 360 | BDBR: -18.20% -> -16.84% BDBR vs LDMIC | Mixed |
Closest papers by meaning
- 3D-LMVIC: Learning-based Multi-View Image Compression with 3D Gaussian Geometric Priors (ICML 2025 · similarity 0.85)
- When MLLMs Meets Compression Distortion: A Coding Paradigm Tailored to MLLMs (ICLR 2026 · similarity 0.85)
- ZPressor: Bottleneck-Aware Compression for Scalable Feed-Forward 3DGS (NeurIPS 2025 · similarity 0.84)
- CAT-3DGS: A Context-Adaptive Triplane Approach to Rate-Distortion-Optimized 3DGS Compression (ICLR 2025 · similarity 0.84)
- 4DGC: Rate-Aware 4D Gaussian Compression for Efficient Streamable Free-Viewpoint Video (CVPR 2025 · similarity 0.83)
- Ultra Lowrate Image Compression with Semantic Residual Coding and Compression-aware Diffusion (ICML 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.
