Research map / Point clouds / Point cloud classification
A Cloud 3D Dataset and Application-Specific Learned Image Compression in Cloud 3D
Tianyi Liu, Sen He, Vinodh Kumaran Jayakumar, Wei Wang
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 |
|---|---|---|---|
| train application-specific models instead of one general model (ablation) | Coarse-to-Fine (large) · Cloud 3D gaming images | PSNR at same bitrate: 2 to 4dB | Improved |
| slim and app-specific model on Imhotep dataset | Coarse-to-Fine · Imhotep | bitrate at 36 dB PSNR: saved about 0.3 bpp with a PSNR of 36dB | Improved |
| reduce channels from Large to Medium (ablation) | app-specific Coarse-to-Fine · Cloud 3D gaming images | compression time (s): 4.16 -> 2.24 | Improved |
| reduce channels from Medium to Small (ablation) | app-specific Coarse-to-Fine · Cloud 3D gaming images | compression time (s): 2.24 -> 1.63 | Improved |
| reduce channels from Small to XSmall (ablation) | app-specific Coarse-to-Fine · Cloud 3D gaming images | rate-distortion | Hurt |
| reduce channels from Small to XSmall (ablation) | app-specific Coarse-to-Fine · Cloud 3D gaming images | compression time (s): 1.63 -> 1.43 | Improved |
| add Model-Task Balance (larger input tiles) to Small model (ablation) | Small model · Cloud 3D gaming images | compression time (s): 1.63 -> 0.93 | Improved |
| add Model-Task Balance (larger input tiles) to Small model (ablation) | Small model · Cloud 3D gaming images | decompression time (s): 2.60 -> 1.75 | Improved |
| prune the last hyperprior layer (ha2/hs2) (ablation) | Small model with M-T-Balance · Cloud 3D gaming images | compression time (s): 0.93 -> 0.74 | Improved |
| prune the last hyperprior layer (ha2/hs2) (ablation) | Small model with M-T-Balance · Cloud 3D gaming images | rate-distortion | No clear effect |
| slim application-specific model replaces Coarse-to-Fine | Coarse-to-Fine · Cloud 3D gaming images | compression time (s): 6.52 -> 0.74 | Improved |
| slim application-specific model replaces Coarse-to-Fine | Coarse-to-Fine · Cloud 3D gaming images | decompression time (s): 7.72 -> 1.64 | Improved |
Closest papers by meaning
- ZPressor: Bottleneck-Aware Compression for Scalable Feed-Forward 3DGS (NeurIPS 2025 · similarity 0.85)
- 3D-LMVIC: Learning-based Multi-View Image Compression with 3D Gaussian Geometric Priors (ICML 2025 · similarity 0.84)
- Compressed 3D Gaussian Splatting for Accelerated Novel View Synthesis (CVPR 2024 · similarity 0.84)
- CAT-3DGS: A Context-Adaptive Triplane Approach to Rate-Distortion-Optimized 3DGS Compression (ICLR 2025 · similarity 0.83)
- UniPCGC: Towards Practical Point Cloud Geometry Compression via an Efficient Unified Approach (AAAI 2025 · similarity 0.83)
- Compressing Streamable Free-Viewpoint Videos to 0.1 MB per Frame (AAAI 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.
