AI Paper Atlas

Research map / Point clouds / Shape and surface completion

R3-PCQA: Ray-Reprojection-Reinforcement for No-Reference 3D Point Cloud Quality Assessment

Junhyuk Seo, Sanghyuk Seo, Dawoon Kim, Heeseok Oh

CVPR 2026 · Point clouds › Shape and surface completion · approach: Point cloud upsampling and normals

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 changedWhereMeasuredEffect
add 3D point cloud input to 2D image input (ablation)R3-PCQA · WPCSRCC: 0.8868 -> 0.9017Improved
add 2D image input to point cloud input (ablation)R3-PCQA · WPCSRCC: 0.7334 -> 0.9017Improved
add local view quality loss, then policy loss, to global loss (ablation)R3-PCQA · WPCSRCC: 0.8644 -> 0.9017Improved
use RL-selected single subcloud instead of fusing all 9 subclouds (910 points) (ablation)R3-PCQA · WPCSRCC: 0.8846 -> 0.9017Improved
use RL-selected subcloud instead of fusing all 9 full subclouds (8192 points) (ablation)R3-PCQA · WPCGFLOPs: 479.32 -> 232.36Improved
use RL-selected subcloud instead of Fusion-910 on compute (ablation)R3-PCQA · WPCGFLOPs: 182.93 -> 232.36Hurt
train on WPC and test on other datasetsR3-PCQA vs MM-PCQA · Cross-dataset PLCCPLCC: 0.861 -> 0.920 (WPC2.0); 0.682 -> 0.721 (SJTU)Improved
train on the small SJTU dataset and test on WPCR3-PCQA vs MM-PCQA · Cross-dataset PLCCPLCC: 0.351 -> 0.273Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy