AI Paper Atlas

Research map / Point clouds / Point cloud classification

Mitigating Geometric Degradation in Fast DownSampling via FastAdapter for Point Cloud Segmentation

Shuofeng Sun, Haibin Yan

ICCV 2025 · Point clouds › Point cloud classification · approach: Efficient point sampling architectures · 2 citations (OpenAlex)

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
replace FPS with random sampling without adaptationPointMetaBase-XL · ScanNet valmIoU: 71.8 -> 68.9Hurt
add FastAdapter and fine-tune with random samplingPointMetaBase-XL · ScanNet valmIoU: 68.9 -> 71.3Improved
random sampling plus FastAdapter instead of FPSPointMetaBase-L · ScanNet valthroughput: 7.3 -> 23 ins./sec. (+315%)Improved
embed FastAdapter into models already using fast downsamplingRandLA · ScanNet valmIoU: 69.0 -> 70.0Improved
embed FastAdapter into PTV3 with grid samplingPTV3 · ScanNet valmIoU: 77.5 -> 78.0Improved
full fine-tuning instead of FastAdapter (ablation)PointMetaBase-XL RS · S3DIS Area-5mIoU: 71.22 -> 69.42Hurt
add geometry-aware aggregation instead of plain averaging into anchors (ablation)FastAdapter P2A · S3DIS Area-5mIoU: 70.12 -> 70.48Improved
add cross-layer aggregator to geometry-aware aggregation (ablation)FastAdapter P2A · S3DIS Area-5mIoU: 70.48 -> 70.85Improved
add offset embedding and geometric weight instead of adding raw anchor features (ablation)FastAdapter A2P · S3DIS Area-5mIoU: 69.83 -> 71.22Improved
add feature distillation from the FPS model (lambda 10 vs 0) (ablation)FastAdapter · S3DIS Area-5mIoU: 70.51 -> 71.22Improved
increase anchor points beyond 100 (ablation)FastAdapter · S3DIS Area-5mIoU: 71.22 (100) vs 70.92 (250)No clear effect
add FastAdapter to stabilize random sampling across seedsPointMetaBase-XL · S3DIS Area-5mIoU variance: ±1.1 -> ±0.3Improved

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