AI Paper Atlas

Research map / Point clouds / Point cloud classification

CP3: Channel Pruning Plug-In for Point-Based Networks

Yaomin Huang, Ning Liu, Zhengping Che, Zhiyuan Xu, Chaomin Shen, Yaxin Peng, Guixu Zhang, Xinmei Liu, Feifei Feng, Jian Tang

CVPR 2023 · Point clouds › Point cloud classification · approach: Efficient point sampling architectures · 21 citations (OpenAlex)

Read the paper

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 CP³ to HRank channel pruningPointNet++ · ModelNet40 testOA: 92.59 -> 92.95Improved
Add CP³ to ResRep channel pruningPointNeXt-S (C=64) · ScanObjectNN testOA: 86.58 -> 88.52Improved
Add CP³ to CHIP channel pruningPointNeXt-B · S3DIS Area-5mIoU: 66.57 -> 66.80Improved
Add CP³ to ResRep channel pruningVoteNet · ScanNetV2mAP@0.25: 62.45 -> 63.92Improved
Add CP³ to ResRep channel pruningVoteNet · SUN RGB-DmAP@0.50: 36.80 -> 37.37Improved
Add CP³ to ResRep channel pruningGroupFree3D · ScanNetV2mAP@0.25: 67.21 -> 68.57Improved
Add knowledge recycling (KR) without CE (ablation)HRank-pruned PointNeXt-S (C=32) · ScanObjectNN test, FLOPs drop 75%OA: 84.79 -> 85.63Improved
Add coordinate enhancement (CE) without KR (ablation)HRank-pruned PointNeXt-S (C=32) · ScanObjectNN test, FLOPs drop 75%OA: 84.79 -> 85.11Improved
Add CE and KR together (ablation)HRank-pruned PointNeXt-S (C=32) · ScanObjectNN test, FLOPs drop 75%OA: 84.79 -> 86.63Improved
Add knowledge recycling (KR) without CE (ablation)HRank-pruned PointNeXt-S (C=32) · ScanObjectNN test, FLOPs drop 90%OA: 81.33 -> 83.66Improved
Add coordinate enhancement (CE) without KR (ablation)HRank-pruned PointNeXt-S (C=32) · ScanObjectNN test, FLOPs drop 90%OA: 81.33 -> 83.10Improved
Add CE and KR together (ablation)HRank-pruned PointNeXt-S (C=32) · ScanObjectNN test, FLOPs drop 90%OA: 81.33 -> 84.83Improved

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