AI Paper Atlas

Research map / Point clouds / Segmentation and detection

An MIL-Derived Transformer for Weakly Supervised Point Cloud Segmentation

Cheng-Kun Yang, Ji-Jia Wu, Kai-Syun Chen, Yung-Yu Chuang, Yen-Yu Lin

CVPR 2022 · Point clouds › Segmentation and detection · approach: Point cloud semantic segmentation · 56 citations (OpenAlex)

Check 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
Use the proposed method instead of WYPR under scene-level supervisionWeakly supervised point-cloud segmentation · S3DIS scene-levelmIoU: 22.3 -> 12.9Hurt
Use the proposed method instead of MPRM under subcloud-level supervisionWeakly supervised point-cloud segmentation · ScanNetmIoU: 43.2 -> 47.4; 41.1 -> 45.8Improved
Use the proposed method instead of OTOC† with 20 labeled pointsWeakly supervised point-cloud segmentation · ScanNet testmIoU: 55.1 -> 54.4Hurt
Use the proposed method instead of PSD with 0.02% labeled pointsWeakly supervised point-cloud segmentation · S3DIS testmIoU: 48.2 -> 51.4Improved
Add the MIL-derived transformer loss (ablation)CAM classification baseline · ScanNet and S3DIS, sparse labelsmIoU: 52.3 -> 55.4; 46.3 -> 49.1Improved
Add cross-scale consistency to the MIL model (ablation)MIL-derived transformer · ScanNet and S3DIS, sparse labelsmIoU: 55.4 -> 55.9; 49.1 -> 49.6Improved
Add AdaGWP to MIL and cross-scale consistency (ablation)MIL-derived transformer · ScanNet and S3DIS, sparse labelsmIoU: 55.9 -> 57.8; 49.6 -> 51.4Improved
Add an ordinary transformer encoder without the MIL formulation (ablation)CAM classification baseline · ScanNet and S3DIS, sparse labelsmIoU: 52.3 -> 55.0; 46.3 -> 47.9Improved
Use only positive MIL bags rather than both bag types (ablation)MIL-derived transformer · ScanNet and S3DIS, sparse labelsmIoU: 57.8 -> 57.4; 51.4 -> 49.1Hurt
Increase the relative weight of negative bags (α=0.3, β=0.7) (ablation)MIL-derived transformer · ScanNet and S3DIS, sparse labelsmIoU: 57.8 -> 57.1; 51.4 -> 51.2Hurt
Use only negative MIL bags rather than both bag types (ablation)MIL-derived transformer · ScanNet and S3DIS, sparse labelsmIoU: 57.8 -> 56.9; 51.4 -> 48.9Hurt
Increase the cross-scale sampling ratio from γ=0.25 to γ=0.5 (ablation)MIL-derived transformer with cross-scale consistency · ScanNet, sparse labelsmIoU: 55.5 -> 56.4Improved

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